id stringlengths 36 36 | tier stringclasses 1
value | github_login stringclasses 179
values | created_at stringlengths 32 32 | status stringclasses 2
values | sha256 stringlengths 64 64 | source_bytes int64 341 260k | leaderboard_rank int64 | source large_stringlengths 341 260k | metadata_json large_stringlengths 645 734 | result_json large_stringlengths 5 4.8k |
|---|---|---|---|---|---|---|---|---|---|---|
0001d8fa-3147-4fd0-a49d-61614c52d9ac | easy | CodeReclaimers | 2026-08-20 15:21:07.916546+00:00 | succeeded | ecdc5bc3549f21ec4b5703f65802288e92bc05cd2c1587f9346ab4e680aba37d | 42,274 | null | """v39_keyed: keyed-modulus CRT learner — Run A of T1_CERT_PLAN_20260819.
HYPOTHESIS A (alias modulus): h1's displayed N is a stable per-identity
alias phi(M) of a hidden modulus M <= 2^20 from enumerated factor pairs
(<=10-bit primes), while x and y are displayed plainly and the step closes
per prime channel (squarin... | {
"id": "0001d8fa-3147-4fd0-a49d-61614c52d9ac",
"created_at": "2026-08-20 15:21:07.916546+00:00",
"db_md5": "4734e439309d100106007d6b654ccab0",
"submitter": "CodeReclaimers",
"github_login": "CodeReclaimers",
"run_id": "9824eccf-75ad-45fa-86fd-90000b08b1dc",
"tier": "easy",
"dataset_id": "e5",
"status... | {
"score": {
"mean_loss": 1.6747209675214556,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.47916666706403094
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2111226555477876,
"example_count": 600... |
000228af-ba9c-4313-a87f-74121f716157 | easy | liam-gb | 2026-08-11 10:17:03.181174+00:00 | succeeded | a52fd23b420d06c1ea43204544d5e9ffbcb40ea72e22fa95b26f8d1eaa5e17e6 | 12,304 | null | """rns_critical_uni: critical-path loop with uniform per-loop supervision.
Each field value enters as its residue phase on a bank of small odd prime
circles via fixed sinusoidal feature maps over digit significance
(omega[s] = 2*pi*(10^s mod p)/p) — carry-free by construction. Phases are
resolved against fixed unit-ci... | {
"id": "000228af-ba9c-4313-a87f-74121f716157",
"created_at": "2026-08-11 10:17:03.181174+00:00",
"db_md5": "4e4e60e58520faa9b1efeba2156ec5ae",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "15476376-a505-4f85-ac7e-e1c77ec3e5bb",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 5.046521425247192,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0800000000745058
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.268509864807129,
"example_count": 100,
... |
0009d371-ce6b-4cb1-8742-acab0548134c | easy | jordanrubin | 2026-08-06 19:23:31.320469+00:00 | succeeded | 2214473aabeee3e5a3252928aedd50993436405bf900fc579ecc4ef88811f4f4 | 14,788 | null | """Autonomous digit-state recurrent Transformer for repeated modular squaring.
The model is deliberately organized around one learned transition:
p_0 = right_aligned_decimal_digits(x)
p_{k+1} = F_theta(p_k, decimal_digits(N))
The same two-block Transformer cell is applied exactly T times. T is used only
as ... | {
"id": "0009d371-ce6b-4cb1-8742-acab0548134c",
"created_at": "2026-08-06 19:23:31.320469+00:00",
"db_md5": "f917ba4387feee357d1f6822e4a7825f",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "7a653879-d845-4735-89f6-8d24230933d3",
"tier": "easy",
"dataset_id": "e3",
"status": "s... | {
"score": {
"mean_loss": 2.1296750745907396,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0018749999580904841
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1309655856123277,
"example_count": 8... |
000a3457-d237-4459-bed2-b08371ee3d12 | easy | sapient-sapiens | 2026-08-19 19:24:48.910974+00:00 | succeeded | daa4fc1f1821b95b049e62d1602f10121592d66b9efbbb3e78e141bca4403b2b | 33,252 | null | """Depth-quantized recurrence with a mixture-over-depth objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. There is no s... | {
"id": "000a3457-d237-4459-bed2-b08371ee3d12",
"created_at": "2026-08-19 19:24:48.910974+00:00",
"db_md5": "2553f964fd1803486430c2690edf5c49",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "bc701422-a4a4-4e0b-a61a-aa3779c48292",
"tier": "easy",
"dataset_id": "e3",
"status": "su... | {
"score": {
"mean_loss": 2.2663152426947795,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008125
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.276738232607394,
"example_count": 800,
... |
000c7da0-a4e9-4f0d-ac6d-ceba3bbd79e3 | easy | karanganesan | 2026-08-06 05:45:58.399132+00:00 | succeeded | 35bb515be06389766e7968d9741d65099fd39cb897f018b8a7b9b4d9e5188a5a | 25,616 | null | """Parametric looped-transformer family (P1).
One weight-tied transformer block applied k times in latent space. Config
flags cover four P1 families with one file:
looped recall=0 gated=0 tfilm=0 plain weight-tied loop
looped-recall recall=1 re-inject the input embedding each
... | {
"id": "000c7da0-a4e9-4f0d-ac6d-ceba3bbd79e3",
"created_at": "2026-08-06 05:45:58.399132+00:00",
"db_md5": "853fe7df67aa06b7473fbcdd5ee2d312",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "9ce87288-351f-43fc-9286-32c267afb2c7",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 3.7291706800460815,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666716337204
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.6878952980041504,
"example_count": 10... |
0019ad4c-eb58-4195-91c7-8648d5bcbcb6 | easy | ArkinDharawat | 2026-08-15 20:08:59.860054+00:00 | succeeded | b3a3fc7ccd6a67634a1aaadb13d50e511a3f9ba42651a0b7a87cef3c72bf10ef | 7,633 | null | """Universal-Transformer-style looped block, v2: step embed + pos offset + LR schedule + 6 loops.
Extends ``looped_ut_padded`` with four ideas drawn from
Graves 2016 (ACT), Dehghani et al. 2018 (Universal Transformer), and
Merrill & Sabharwal 2025 (log-depth transformers):
1. **Learned per-step (timestep) embedding**... | {
"id": "0019ad4c-eb58-4195-91c7-8648d5bcbcb6",
"created_at": "2026-08-15 20:08:59.860054+00:00",
"db_md5": "df88b48f30c8035a9269455deffdd66b",
"submitter": "Arkin Dharawat",
"github_login": "ArkinDharawat",
"run_id": "d7867b75-a51b-4a0b-8db9-6c11320331f7",
"tier": "easy",
"dataset_id": "e3",
"status"... | {
"score": {
"mean_loss": 2.240777682338874,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009375000055879355
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2808395810221542,
"example_count": 800... |
001d49d5-358e-4e45-9cab-9a698d534eb3 | easy | mnida | 2026-08-25 22:18:55.230889+00:00 | succeeded | a18891d3780d413e19103a57d2fe22a3de5a430300e068f22fb0df54e9f2b708 | 15,662 | null | """Large quadratic looped Transformer with immutable direct-N-blind routes."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_stat... | {
"id": "001d49d5-358e-4e45-9cab-9a698d534eb3",
"created_at": "2026-08-25 22:18:55.230889+00:00",
"db_md5": "f0417c54402832ba8b3df1ddb65117d5",
"submitter": "mnida",
"github_login": "mnida",
"run_id": "aa85254a-c490-4ac1-ac7e-d3f6ccd111a5",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"score": {
"mean_loss": 2.1489941186962733,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0025000000124176342
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1486757325484613,
"example_count": 6... |
001e4315-d76b-43da-8a88-0f8a9d2e3f95 | easy | nikolageorgiev2000 | 2026-08-04 15:19:10.646519+00:00 | succeeded | bcb781bfdf7ba38a0a37eaa6ed064a66f4dfd339aab3b052a335c7c7073403a1 | 26,498 | null | """T-composed categorical-digit reasoner with fully learned pair/reduction maps."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
asser... | {
"id": "001e4315-d76b-43da-8a88-0f8a9d2e3f95",
"created_at": "2026-08-04 15:19:10.646519+00:00",
"db_md5": "f86607b469fd9dd75773b8607f53025a",
"submitter": "Nikola Georgiev",
"github_login": "nikolageorgiev2000",
"run_id": "2e4811a7-b113-43e4-b93f-c7c51913c2bf",
"tier": "easy",
"dataset_id": "e3",
"s... | {
"score": {
"mean_loss": 2.206954932994406,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010625000009313226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1918833108323432,
"example_count": 800... |
00224bd0-73fb-4a40-97f8-a9b07f9da26e | easy | shreyash-chonkie | 2026-08-26 14:24:03.669929+00:00 | succeeded | c55d3c9a0c2334b02bb81847ede9332c2e7c70fa8081fb22b066ce23f236b91d | 11,398 | null | """Ternary register machine with supervised execution prefixes."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_mo... | {
"id": "00224bd0-73fb-4a40-97f8-a9b07f9da26e",
"created_at": "2026-08-26 14:24:03.669929+00:00",
"db_md5": "8181bb7dc94c319424c0d3bdffb13103",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "147e8a08-b666-4c55-847e-74b18780ca5e",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 1.944562554359436,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.054999999701976776
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9016361236572266,
"example_count": 100... |
0029af78-e042-41d6-8ff6-f63b473acf21 | easy | erdavis0 | 2026-08-22 07:43:23.646040+00:00 | succeeded | a1c22e3f3f7475c81762f81ac0379de4cfc7d9c3c89e6889339eb862577f0cc4 | 6,485 | null | """Field crossbar with learned row and column summaries."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
WIDTH = 64
FIELDS = 6
HID... | {
"id": "0029af78-e042-41d6-8ff6-f63b473acf21",
"created_at": "2026-08-22 07:43:23.646040+00:00",
"db_md5": "007d6bd26662c86be5ce3a5c495153a4",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "a13c3c44-4353-46bd-ac60-aed72f096eec",
"tier": "easy",
"dataset_id": "e8",
"status": "succeeded",... | {
"score": {
"mean_loss": 6.961434841156006,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03698752261698246
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.846958160400391,
"example_count": 85,
... |
0031a018-34be-4ef6-84f0-84d7b7bb28e6 | easy | isaac0804 | 2026-08-10 21:43:37.056044+00:00 | succeeded | 5a296f34bfec92574192274ead57580898d5e214659006262dccab29ab3dcf16 | 7,376 | null | """Clean loop (no damped gate, no step-embedding) + general learnable
scratchpad registers, carried through the loop's persistent state.
Distinct from two things already tried and closed:
- `submissions/scratchpad_register`: the same 16-general-register idea, but
on the plain untied 8-layer architecture (E3: 0.88%, ... | {
"id": "0031a018-34be-4ef6-84f0-84d7b7bb28e6",
"created_at": "2026-08-10 21:43:37.056044+00:00",
"db_md5": "8ea092d16b85bdf5174b7161cd78b953",
"submitter": "Isaac Yong",
"github_login": "isaac0804",
"run_id": "63c63a6b-9bf7-4ec0-892c-349a11721d2a",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 4.23493684525125,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012500000121071934
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.400024137829952,
"example_count": 800,
... |
004f7670-b2fa-4551-83c3-652c5afe9f1a | easy | lzy54 | 2026-08-27 05:36:54.828679+00:00 | succeeded | cff143bebd9be0f8c903f1768d214d90c30d347351c5be2e8e6bbd3ee2c871fe | 6,885 | null | """B2 plus weak supervision at the sampled endpoint's penultimate state."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_... | {
"id": "004f7670-b2fa-4551-83c3-652c5afe9f1a",
"created_at": "2026-08-27 05:36:54.828679+00:00",
"db_md5": "46654244fbe3f2a394a320430400fefb",
"submitter": "lzy54",
"github_login": "lzy54",
"run_id": "44205fb6-0c77-433e-8a52-f76558d4a184",
"tier": "easy",
"dataset_id": "e6",
"status": "succeeded",
... | {
"score": {
"mean_loss": 2.6724324226379395,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.043468470685184
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.531221628189087,
"example_count": 60,
... |
00502292-2ea4-40fe-839f-e64ce37a0e54 | easy | oupadhyay | 2026-08-28 02:30:14.767682+00:00 | succeeded | d9fe46d380639e44f497127eb4fcfbb43643dd68644e1e8f3155f85bad84b9cc | 8,697 | null | """N-conditioned factorized ordered digit-pair context with a tied reducer."""
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import OptimizerBundle, Submission, assert_model_state
PAD, N, X, T, ANS, DIGIT = 0, 2, 3, 4, 5, 7
D, PLACES, MICROPHASES, MAX_STEPS = 64, 4, 1, 64
STATE_ELEM... | {
"id": "00502292-2ea4-40fe-839f-e64ce37a0e54",
"created_at": "2026-08-28 02:30:14.767682+00:00",
"db_md5": "98a660c64e7ac2316a3d496ca06d2202",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "885db408-3b62-4d18-b83d-308cc1927192",
"tier": "easy",
"dataset_id": "e6",
"status": "s... | {
"score": {
"mean_loss": 1.964000940322876,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.043468470685184
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9523199796676636,
"example_count": 60,
... |
0050263d-d280-4bbf-acbe-92e498549808 | easy | alirezashirvani-jr | 2026-08-31 00:51:59.583576+00:00 | succeeded | d30c15c4b819074a695cfe7f9ab43fd9530cdf8d2f612f55533983ffcf4e7dfb | 22,445 | null | from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
PERIOD_MINIMUM = 2
PERIOD_MAXIMUM = 96
EXP... | {
"id": "0050263d-d280-4bbf-acbe-92e498549808",
"created_at": "2026-08-31 00:51:59.583576+00:00",
"db_md5": "a726edbe0af9de151bd033a13eebcee4",
"submitter": "alirezashirvani-jr",
"github_login": "alirezashirvani-jr",
"run_id": "32ac725f-0540-45ca-a521-bae8ea946d2b",
"tier": "easy",
"dataset_id": "e9",
... | {
"score": {
"mean_loss": 0.0,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.0,
"example_count": 90,
"exact_accuracy": 1.0,
"c... |
00512563-3290-4db6-8889-44c706fa7f91 | easy | DDanlov | 2026-08-19 20:24:20.501996+00:00 | succeeded | 6279b4a3803a8a8c508b6fa8486b0772665676a329ba166b1310256f3d7077ea | 17,727 | null | """
Dynamic Submission for One Layer Deeper API Benchmark
Config: dim=2240, num_heads=32, d_ff_mult=4, fixed_trec=224
"""
from __future__ import annotations
import math
from typing import Any, Optional, Tuple
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
... | {
"id": "00512563-3290-4db6-8889-44c706fa7f91",
"created_at": "2026-08-19 20:24:20.501996+00:00",
"db_md5": "837bab05d875c88343d5f22a5ca8c7e3",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "eff5e4a9-a496-4e35-a364-fe4d15c7377d",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.5423725202964977,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006250000024835269
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.650904630866286,
"example_count": 600... |
00512b6e-a424-466e-87f6-841f8951dbb4 | easy | DDanlov | 2026-08-20 22:41:31.203907+00:00 | failed | 9c0511bb8650a59808df8f3a8b68beada95b0d4125faedacb8310ec54de0db87 | 21,251 | null | from __future__ import annotations
import math
import time
from typing import Any, Optional, Tuple
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from benchmark.api import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
except Imp... | {
"id": "00512b6e-a424-466e-87f6-841f8951dbb4",
"created_at": "2026-08-20 22:41:31.203907+00:00",
"db_md5": "554a5425b8279bdb1c27f8908c3466ce",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "f4ecd7be-f4f7-46fd-bc5f-72c7489be558",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",
... | null
|
0051c806-aac6-4991-ba26-9dda12e8fd28 | easy | jordanrubin | 2026-08-07 06:04:08.286696+00:00 | succeeded | 6f17eb27b2bb1fbb5891f0392535694f326c447b00cd9f37136d01d78d2e30ec | 16,908 | null | """Autonomous digit-state recurrent Transformer for repeated modular squaring.
The model is deliberately organized around one learned transition:
p_0 = right_aligned_decimal_digits(x)
p_{k+1} = F_theta(p_k, decimal_digits(N))
The same two-block Transformer cell is applied exactly T times. T is used only
as ... | {
"id": "0051c806-aac6-4991-ba26-9dda12e8fd28",
"created_at": "2026-08-07 06:04:08.286696+00:00",
"db_md5": "378594dd4459a9c7efa73ec8d9805eae",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "2464da5b-7502-42a9-91de-d8972268fb48",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 4.1554332607364906,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012083333063249786
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.212000492308706,
"example_count": 600... |
00575b42-67c7-437a-9663-c0daa6bf8b21 | easy | AdityaVG13 | 2026-08-15 11:10:01.532762+00:00 | succeeded | c804653af9f9ffa200b4c1da8efdb3beb92f7dc867f10dd7e6fe23599b08ba27 | 109,418 | null | """AEGIS V12 v12_fc_carry20: frontier-carry with SOFT_CARRY=0.20.
One Layer Deeper submission candidate. All task transitions remain learned
end-to-end. The deterministic shell only identifies marker-delimited roles,
constructs a finite semantic tape, schedules learned interval programs, and
applies a trivalent Safe/U... | {
"id": "00575b42-67c7-437a-9663-c0daa6bf8b21",
"created_at": "2026-08-15 11:10:01.532762+00:00",
"db_md5": "9a7f754c6d74c7f68a4c8989b264cb9c",
"submitter": "AdityaG",
"github_login": "AdityaVG13",
"run_id": "7a80020e-8cb6-4baa-bf87-2a3c088de17f",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 2.0099394023418427,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.7166666388511658
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.5161845088005066,
"example_count": 100,... |
005b8db6-601a-4128-bc1f-b32bf877932f | easy | sapient-sapiens | 2026-08-09 21:35:10.493743+00:00 | succeeded | 5bb4e61f0c3687019fbff8d90d6b56ec8467719a08ba00c1fe0e69800328728e | 15,567 | null | """T-independent categorical recurrence with learned latent-depth selection."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
Toke... | {
"id": "005b8db6-601a-4128-bc1f-b32bf877932f",
"created_at": "2026-08-09 21:35:10.493743+00:00",
"db_md5": "d5518602a2c831b1a0df7d2053fd26b5",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "880617d4-3826-403a-8ead-5eeb5f75bdc4",
"tier": "easy",
"dataset_id": "e4",
"status": "su... | {
"score": {
"mean_loss": 2.131069280855379,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0017708333333333335
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.138858349809138,
"example_count": 120... |
005d0d72-a3a8-4fef-8e99-c74967cb28b9 | easy | EthanBnntt | 2026-08-13 17:09:00.770054+00:00 | succeeded | 3f1efd42b4c527f32824c0ae4e467c16b3ede032a10202d0dfdb94a6348d1253 | 12,640 | null | """Variant D — supervision/curriculum evolution of the shared-weight
recurrent-refinement transformer (HRM-style).
Same lawful, generic architecture as the known-good baseline: one shared block
repeated over the whole sequence so the network can perform substantially deeper
serial computation, plus deep supervision ov... | {
"id": "005d0d72-a3a8-4fef-8e99-c74967cb28b9",
"created_at": "2026-08-13 17:09:00.770054+00:00",
"db_md5": "00ba39f8cd4f85a0ccb4e913d350b8d8",
"submitter": "Bennett",
"github_login": "EthanBnntt",
"run_id": "3821bee9-58c0-496f-a1cc-1107426e568d",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.1574717715222445,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009166666679084301
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1620072078277177,
"example_count": 60... |
0066c014-db68-4577-893e-a0652f12f17f | easy | sapient-sapiens | 2026-08-19 18:44:01.906411+00:00 | succeeded | 2f306ce43511ab5e1ab51d0e20ce3e35ae571f6a2d5cbff7e231f9b6deec6c03 | 32,714 | null | """Depth-quantized recurrence with a mixture-over-depth objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. There is no s... | {
"id": "0066c014-db68-4577-893e-a0652f12f17f",
"created_at": "2026-08-19 18:44:01.906411+00:00",
"db_md5": "d2fe6eaa14bdd8d94eaf90a2e2b79466",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "8777e4b0-272b-4b0d-8e16-0ab79c428260",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.2068501783150585,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00625
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2033536388735064,
"example_count": 600,
... |
006fce9c-39b8-4070-9750-87b5646229d7 | easy | DDanlov | 2026-08-29 13:56:14.982895+00:00 | succeeded | 4312e34c5437adbb0bc116a741abec3d64983625b501fe931039ff8a92fede21 | 24,013 | null | """
trial_w14_l5_d512_k8.py
W14: L=5, D=512, H=16, d_ffn=2048, K=8, dt=0.08, mu=0.85, lr=0.003, freq=8
"""
import math
import time
import contextlib
from dataclasses import dataclass
from typing import Optional, Tuple, Dict, Any, List, Union
import torch
import torch.nn as nn
import torch.nn.functional as F
try:
... | {
"id": "006fce9c-39b8-4070-9750-87b5646229d7",
"created_at": "2026-08-29 13:56:14.982895+00:00",
"db_md5": "d12d6faf4272ab464c3a5c8dcbc6c0ed",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "1ea823b0-ac2e-4e9e-a8dd-4297304fb8b3",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.2805421419198795,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2819681603277746,
"example_count": 600,... |
0078ca63-9c29-4808-8e69-f7e4fad1ec51 | easy | 0Chris5R | 2026-08-07 12:40:24.990343+00:00 | succeeded | a3cf17ab079ff0003d74dceb0d39327c4cb33d1bc3e2046eb357cd0eaa7ec94b | 20,839 | null | """Shared-digit Neural GPU with a compute-matched operator curriculum.
The model aligns decimal digits by significance, forms a learned low-rank
second-order feature map of the current state, and evolves a two-dimensional
workspace with the original convolutional-GRU equations. Each arithmetic
transition applies the ... | {
"id": "0078ca63-9c29-4808-8e69-f7e4fad1ec51",
"created_at": "2026-08-07 12:40:24.990343+00:00",
"db_md5": "bb5353e85ef093c19eccba17c7f3c5c3",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "21f793cd-864d-49d2-8b5f-e9bd582c7e4b",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.7946976789304587,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018333333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.7421974291181352,
"example_count": 60... |
007cd212-3bfe-4a54-8ee6-9a171ba7a623 | easy | khushidahi | 2026-08-12 20:29:28.932871+00:00 | succeeded | 31268bc88521e6039a7f218f66834dba2ac9a4e22d808888d8828da9463de5af | 24,153 | null | """R53 learned digit-pair table with directional carry/reduction scans.
The model learns a full 10x10 pair embedding table from final-label training.
For the diagonal variants, pair feature (i,j) is routed to decimal place i+j.
No multiplication values, carry rules, modular-reduction rules, generated
examples, or inte... | {
"id": "007cd212-3bfe-4a54-8ee6-9a171ba7a623",
"created_at": "2026-08-12 20:29:28.932871+00:00",
"db_md5": "80673e16e0f61efcad83531f31d1fe7f",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "3af4a65d-b2ef-49ca-8abf-4c2e1475af08",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.8364956378936768,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009166666772216558
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.877579927444458,
"example_count": 600... |
00819f0f-3b08-4b2d-a763-9934b1232eb4 | easy | erdavis0 | 2026-08-19 07:38:17.389146+00:00 | succeeded | 6281314aefd8eb1f96778ae544917e728a8865c1dafcee55631d6d26f261dd6b | 8,023 | null | """Parser-free neural cellular transducer frozen at the 100-update checkpoint."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
To... | {
"id": "00819f0f-3b08-4b2d-a763-9934b1232eb4",
"created_at": "2026-08-19 07:38:17.389146+00:00",
"db_md5": "2765ec39948d9079ec124508edb3e825",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "600e3dba-f913-4754-ac49-93b1260b96bb",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",... | {
"score": {
"mean_loss": 1.8256409764289856,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666716337204
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.7528597116470337,
"example_count": 10... |
00838ab7-f987-49ea-986d-63b6049e1a67 | easy | lpbb | 2026-08-07 18:54:14.885410+00:00 | succeeded | 1577b3efd1dce30988090b1309386e8a095f5088e780d1f367fc5fbb7962b264 | 24,045 | null | """VARIANT R2b `fourier` — representation-prior axis (phase bank).
Where the campaign stands: memorization cannot be blocked (d=64 with 106k
params still memorizes 3200 rows in 45 s), delayed into irrelevance
(wd 1.0 at ~1000 full-batch epochs), or taxed away (softcap+smoothing
bounds confidence, map unchanged). The ... | {
"id": "00838ab7-f987-49ea-986d-63b6049e1a67",
"created_at": "2026-08-07 18:54:14.885410+00:00",
"db_md5": "73f171ca1d959ec9da637787c40d5cde",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "d122b03c-6b7f-40d4-abce-3c2bb397875d",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 4.218889554529824,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.020000000027939675
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.294286633420502,
"example_count": 800,... |
0086502e-95ca-46b4-a58f-f8e996c577ac | easy | sapient-sapiens | 2026-08-07 07:39:00.787589+00:00 | succeeded | f2f6b6d85f8cd4f74322148c8fe6180ac151c635eb26969d3725a504f3c5f284 | 11,764 | null | """Digit-space recurrent operator for repeated modular squaring.
Nothing here converts a token into a decimal number. Digits stay as embeddings,
so what a digit is worth, and every carry and reduction, is learned. That also
sidesteps a hard numeric wall: at Hard's modulus scale `x^2` runs past the 53-bit
exact-integer... | {
"id": "0086502e-95ca-46b4-a58f-f8e996c577ac",
"created_at": "2026-08-07 07:39:00.787589+00:00",
"db_md5": "63dae53794d440f68032bbd3d18fa7bc",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "8aac6588-eb33-4a50-a438-a6a15768dbe4",
"tier": "easy",
"dataset_id": "e1",
"status": "su... | {
"score": {
"mean_loss": 3.6392561607527316,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.2383333338300387
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.01665735244751,
"example_count": 100,
... |
008ada70-9917-41c7-98ae-be6e4e5583c8 | easy | repst | 2026-08-25 01:30:12.783641+00:00 | succeeded | 5fa4f569d54314c7eb4d8a9b2b1b740d4fefa922ca83be75ddb5c157ebc6f886 | 50,446 | null | # R657: accelerator-only delta from frozen R655.
# Deep proposal syntax, semantics, scores, posterior learning, MDL, parameters,
# and pure-neural inference are unchanged. Training compiles proposed polynomial
# denotations into one dense coefficient tensor and contracts shared monomial
# caches, replacing R655's many ... | {
"id": "008ada70-9917-41c7-98ae-be6e4e5583c8",
"created_at": "2026-08-25 01:30:12.783641+00:00",
"db_md5": "016311e89cb5d74685a2145a41e1c2c5",
"submitter": "Ryan Epstein",
"github_login": "repst",
"run_id": "4a958631-2624-42cc-bf35-a00e90d3a0ba",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 0.2254623125689961,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.22665289785261913,
"example_count": 600,
"ex... |
00934db9-fde8-43b7-8786-1f53130d6252 | easy | rishivg | 2026-08-04 09:29:15.456917+00:00 | succeeded | b35bd60a4eefe9eaa88af1035fd71eaf7f05dd96f6480bc1310654bdba9b93eb | 131,327 | null | """Looped transformer with place-value coordinates, Muon, and a clock-paced LR.
Design notes, in rough order of expected value:
1. Weight-tied recurrence with input injection. The task is function
composition of depth ``T``, so one block applied ``LOOPS`` times is a much
better prior than a single feed-forw... | {
"id": "00934db9-fde8-43b7-8786-1f53130d6252",
"created_at": "2026-08-04 09:29:15.456917+00:00",
"db_md5": "5cdc8956d07eb3fe563d5d24dfbe06ee",
"submitter": "Rishi Gottumukkala",
"github_login": "rishivg",
"run_id": "b04faeee-8c62-46cf-965b-f879fa4add53",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 15.318617820739746,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.013750000391155481
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 15.325847625732422,
"example_count": 60... |
009c34c6-b400-417d-8dbd-715d842d9792 | easy | sapient-sapiens | 2026-08-15 20:12:02.788897+00:00 | succeeded | 469ae48173121f8a328b5c1f44d0bff0e2f098d243dcd227975cee12399b6829 | 17,772 | null | """T-only adaptive chunk routing over a learned shared operator.
The controller sees only the categorical tokens inside the T prompt field and
is trained solely by final-answer loss. It never parses T numerically and T is
not used as an intermediate supervision target. Inputs use learned additive
token, semantic-gro... | {
"id": "009c34c6-b400-417d-8dbd-715d842d9792",
"created_at": "2026-08-15 20:12:02.788897+00:00",
"db_md5": "38cda1d694d318221e86f82dd23bbc39",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "c6c81ece-6618-45b7-b9bd-a4aa3d4f04a1",
"tier": "easy",
"dataset_id": "e9",
"status": "su... | {
"score": {
"mean_loss": 7.7519566069922465,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05000000203649203
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.106731414794922,
"example_count": 90,
... |
009ce33c-ab1d-4cb6-be74-ac24758c7821 | easy | lpbb | 2026-08-14 09:41:56.835734+00:00 | succeeded | 4aebad7a5eb227f833ac85d3524bfa29eb52689d4caa5da2675c222c06f349e0 | 7,100 | null | """S4 — S2 with digit embeddings initialized as multi-frequency circle features
(the full Fourier basis of Z_10), tilted into the full width by a fixed random
rotation. Everything stays learnable; only starting values change."""
from __future__ import annotations
import math
import torch
import torch.nn.functional a... | {
"id": "009ce33c-ab1d-4cb6-be74-ac24758c7821",
"created_at": "2026-08-14 09:41:56.835734+00:00",
"db_md5": "23e774c886ac7e95e81d22046bfbb1e2",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "8e772f95-63a2-41d4-82ad-6edf02d76b87",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 2.3783113956451416,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005624999874271452
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.434177875518799,
"example_count": 800... |
009d4478-62a8-4b48-8a63-d30fa3ec3c62 | easy | viridale | 2026-08-05 22:19:23.787051+00:00 | succeeded | 842b56e00faf472d166b7c1de91552d6d15e3cc7d3322eec3ec6956a653ed4e3 | 9,914 | null | """Clean v177 with sequence-global contrast on provided final strings.
hypothesis: independent token CE permits mutually inconsistent target-slot features
and the stable 8.33% basin; aligning one generic prompt code to one learned code
for the supplied whole final string may identify exact row-level prediction... | {
"id": "009d4478-62a8-4b48-8a63-d30fa3ec3c62",
"created_at": "2026-08-05 22:19:23.787051+00:00",
"db_md5": "e7fa47a122e116097c8cbf2de78eb1bd",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "d3f7bbc5-6126-43c0-9de8-1d7afd283753",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 2.1813634634017944,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0833333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1977760791778564,
"example_count": 100,... |
00c79a75-1621-4358-9722-ddcad009b121 | easy | k-penchev | 2026-08-09 17:05:56.667408+00:00 | succeeded | 7bf30e19d90e7f6cadb4ad67d0b2cd51634a4fd2d29a507be6f840a0dd78f8c1 | 3,628 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "00c79a75-1621-4358-9722-ddcad009b121",
"created_at": "2026-08-09 17:05:56.667408+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "Kaloyan Penchev",
"github_login": "k-penchev",
"run_id": "488204e4-d5ec-42a7-a8d4-21431672a633",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 3.0323007106781006,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.031759262084961,
"example_count": 100,
"exac... |
00d4bc80-2479-4df8-8390-a815003fa1a1 | easy | oupadhyay | 2026-08-25 04:04:38.200139+00:00 | succeeded | 4f61c1933d2aef421b76201aa4c7649fae181badb245d4d2fff2ca1fe0fe5a7f | 5,478 | null | """Shared query-pooled relation bank with an LSD-first recurrent decoder."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import OptimizerBundle,Submission,TokenLossBatch,assert_model_state
W,D,R,A,V,H,P=4,32,32,4,8,64,8
class C:
def __... | {
"id": "00d4bc80-2479-4df8-8390-a815003fa1a1",
"created_at": "2026-08-25 04:04:38.200139+00:00",
"db_md5": "deabba845c27c69069658e6e37015a2c",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "e0a237bb-02c2-406b-8222-d5d0e1fb7e64",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.202196311359886,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01041666670391957
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.158911383205465,
"example_count": 600,
... |
00dc7ef3-b312-4245-9a72-b412b2288417 | easy | ehonig | 2026-08-14 05:59:29.602655+00:00 | succeeded | f5ef178d1293e0fe1dacdd38f3bff3a79295b353edeb066061ad00e0be24d0e4 | 52,555 | null | """DigitLoop — a place-value recurrent model for repeated modular squaring.
The task is ``y = x^(2^T) mod N`` with the answer read back tail aligned, one
decimal place value per token position from the right. A stack of attention and
MLP layers has no primitive for the two operations this needs — a partial-product
co... | {
"id": "00dc7ef3-b312-4245-9a72-b412b2288417",
"created_at": "2026-08-14 05:59:29.602655+00:00",
"db_md5": "07299484734830f18ebf65bfe022434f",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "46c1be7f-2531-4965-bbe4-25093cfeff4d",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 5.111151218414307,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.4000000059604645
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.639484405517578,
"example_count": 100,
... |
00e5f651-de40-4360-afb6-4aabd4fde24a | easy | viridale | 2026-08-09 04:49:49.896370+00:00 | succeeded | 04ffbeaa79110ebcd53413ded3a2c0f3afa7d601262be8f9724ca71adda72b31 | 13,819 | null | """Whole-number-hyperconditioned multiscale categorical transducer.
hypothesis: v392 established that shared multiscale place communication improves E1,
but its transition sees N and X only as separate local digits and exact replicas
transfer 0--2/38 held T1 rows. Encode each complete documented numeral with ... | {
"id": "00e5f651-de40-4360-afb6-4aabd4fde24a",
"created_at": "2026-08-09 04:49:49.896370+00:00",
"db_md5": "7dc352460209ee14fd0c37ed56b2cb1b",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "7a63a3f1-4f7c-4d97-8fe4-51375f63c966",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 2.9095706939697266,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08833333104848862
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.6808300018310547,
"example_count": 100... |
00e7f779-66fa-441b-8ce1-a853c42ad813 | easy | DDanlov | 2026-08-29 14:18:38.183090+00:00 | succeeded | cd25127c8983b41f11d8533eb8021733b039c9d4ea80a4536b0970dfed36dfdc | 27,030 | null | """
trial_l01_l4_d512_k24_rel05.py
L01: L=4, D=512, K_max=24, rel_force=5% (0.05), progress_thresh=0.75, lr=0.0035
"""
import math
import time
import contextlib
from dataclasses import dataclass
from typing import Optional, Tuple, Dict, Any, List, Union
import torch
import torch.nn as nn
import torch.nn.functional as ... | {
"id": "00e7f779-66fa-441b-8ce1-a853c42ad813",
"created_at": "2026-08-29 14:18:38.183090+00:00",
"db_md5": "a965d8986908ca68a0651b1ef87d3d84",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "6088422a-b973-4506-97d4-c457f6a852be",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.281044608209969,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.290864776201847,
"example_count": 600,
... |
00ebdecf-d534-4231-894c-af338367ac2a | easy | pculaf | 2026-08-25 13:58:57.210414+00:00 | succeeded | b3eb3fb8a08cc858183f73e7a647b0710c03cb193362f6b9cce0ea5e9c67ced0 | 8,288 | null | """Four-loop post-RMSNorm Transformer for One Layer Deeper."""
from __future__ import annotations
from collections.abc import Iterable
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_m... | {
"id": "00ebdecf-d534-4231-894c-af338367ac2a",
"created_at": "2026-08-25 13:58:57.210414+00:00",
"db_md5": "06d3398262efcdfa4bf712724e235f4d",
"submitter": "Pavle Culafic",
"github_login": "pculaf",
"run_id": "f5637bdd-d5e1-4724-aea3-302c5b0fa898",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 1.96827894449234,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.029999999329447746
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9192379713058472,
"example_count": 100,... |
00ec5a41-efad-4fd9-ba8e-ab0a6df4707d | easy | sapient-sapiens | 2026-08-13 19:33:24.601085+00:00 | succeeded | 1af41caeca7f0b953375d86448a2dc3fa3b58a4856de0091d9988518997636c2 | 7,592 | null | """E3 fixed-T=2: one encoder plus a four-block tied state operator."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
WIDTH = 1536
HEAD... | {
"id": "00ec5a41-efad-4fd9-ba8e-ab0a6df4707d",
"created_at": "2026-08-13 19:33:24.601085+00:00",
"db_md5": "6b99c158439ca08f557595dd48d03ce4",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "49ebc5ab-7e00-4bcb-86d2-0c8168d73bbe",
"tier": "easy",
"dataset_id": "e3",
"status": "su... | {
"score": {
"mean_loss": 2.2628636361999415,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.259880904746146,
"example_count": 800,
"exa... |
00efb63b-9b48-419b-b67e-90fea5e7702d | easy | DDanlov | 2026-08-25 22:49:50.006367+00:00 | succeeded | 7d0e1c6be5e446d16cace2b6bb901ee37f73458a2c6d2e2655ca433338c6596b | 37,551 | null | """
Architecture 1: Two-Stage Pure Skipless Deliberation Model with Cross-Attention-First Decoder
Unconstrained SwiGLU + CuttingTheSkip Scaled Uniform Orthogonal (SUO) Initialization.
Design Invariants:
1. Deliberation Core: Multi-Head Self-Attention (Mimetic Q-K) + Unconstrained SwiGLU FFN (No Givens, No Rotations).
... | {
"id": "00efb63b-9b48-419b-b67e-90fea5e7702d",
"created_at": "2026-08-25 22:49:50.006367+00:00",
"db_md5": "4e4a5e5048d02751ed1d539a8e8fb678",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "ec6c25ce-c3e9-419f-8576-7f2e0fa10140",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.833213686943054,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8332138061523438,
"example_count": 100,
"exac... |
00f13058-d990-4e75-80fc-634a299c2221 | easy | DDanlov | 2026-08-12 22:43:56.230437+00:00 | failed | 72c8c48930ebc867eb9821a0507429e0fde39f6088ba701138f2069fa1000344 | 27,573 | null | from __future__ import annotations
import math
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
try:
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
except ... | {
"id": "00f13058-d990-4e75-80fc-634a299c2221",
"created_at": "2026-08-12 22:43:56.230437+00:00",
"db_md5": "94c1d8255e7adc66de71752934b725bf",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "a09c6f99-2c6b-43aa-b11d-1e9e179f0ce4",
"tier": "easy",
"dataset_id": "e2",
"status": "failed",
... | null
|
00fd86dc-15b7-44bb-842f-5398deb216b8 | easy | falloficaruss | 2026-08-06 05:31:19.005798+00:00 | succeeded | a12208a58faf462e8b49c8ede2f425d7a5093e2dff25000348e1dd1383606702 | 10,463 | null | """Looped encoder v4 — generalize, don't memorize.
Evidence so far
---------------
m1: digit-prior collapse, score ≈ 0.07%, Max T none.
e1 d54f8d8f (train@8/eval@24): train 95% exact, test 5% — depth mismatch.
e1 70f45c59 (matched depth 12, batch 512): train 97%, test 1.3%, 238 steps/60s
— pure memorization of the 6... | {
"id": "00fd86dc-15b7-44bb-842f-5398deb216b8",
"created_at": "2026-08-06 05:31:19.005798+00:00",
"db_md5": "a6343bb0c851bc57e3e86b21220df86b",
"submitter": "Abhishek Shinde",
"github_login": "falloficaruss",
"run_id": "8f5e2dfd-fd99-49be-adfe-f694753a0cc9",
"tier": "easy",
"dataset_id": "e1",
"status... | {
"score": {
"mean_loss": 4.337295293807983,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.024999999441206455
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.424253940582275,
"example_count": 100,... |
00fdf63b-4428-4e92-89e1-b213222b13ab | easy | EyimofeA | 2026-08-09 23:45:17.284439+00:00 | succeeded | bfd59b43dc6770adaf075f20166470e3a22e9c53fe29db04a1823da72981af35 | 14,063 | null | """Canonical recurrence with one learned local digit-mixing residual.
The mutable register is initialized from x once. Every subsequent application
of the tied cell receives only the current LSD-first digit state and immutable
N digits. Requested T controls only the number of applications. The state
logits are also th... | {
"id": "00fdf63b-4428-4e92-89e1-b213222b13ab",
"created_at": "2026-08-09 23:45:17.284439+00:00",
"db_md5": "e26a4123f5c7ee26953e6b368375147a",
"submitter": "mof",
"github_login": "EyimofeA",
"run_id": "7e80c01c-202e-4ec0-9836-2dfa5cc61ed7",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"score": {
"mean_loss": 2.6529314723033623,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0041666667194416125
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2601940631866455,
"example_count": 6... |
010b59da-27f3-46f2-b289-044d69fad271 | easy | shirvani-jr | 2026-08-29 00:30:53.791792+00:00 | succeeded | e72847ca643d91a3d6cd379d86d25b2971dbeba6724dd85d2510ba9d6a97215c | 52,235 | null | """Adaptive learned discrete recurrent fabric for One Layer Deeper.
A gradient-trained, operation-free recurrent computational machine. The
model learns, from the competition's endpoint supervision: how to parse the
prompt into a bank of categorical digit registers, a reusable transition
circuit applied serially, a l... | {
"id": "010b59da-27f3-46f2-b289-044d69fad271",
"created_at": "2026-08-29 00:30:53.791792+00:00",
"db_md5": "4a5efb78fbbd6b54e1b0674190435b72",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "b555b34b-2f6f-49dc-a018-bbab6bf27d33",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 1.8758695125579834,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666716337204
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8002746105194092,
"example_count": 10... |
010e1c66-34c4-4be6-950b-433a3007bbc6 | easy | oupadhyay | 2026-08-15 21:49:54.288210+00:00 | succeeded | 9ba1ee4f1a22af0eb86b628c4e2054b313693102850704d7749e8127e5dac62a | 3,989 | null | """Conditional optimizer screen: lr10."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import ModelSpec,OptimizerSpec,OptimizerBundle,Submission,assert_model_state
W,D,H=4,64,32
class C:
def __init__(s,vocab_size,max_seq_len):s.vocab_size,s... | {
"id": "010e1c66-34c4-4be6-950b-433a3007bbc6",
"created_at": "2026-08-15 21:49:54.288210+00:00",
"db_md5": "ae1a6184fb19d563c764afa690b280eb",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "896e85c1-828d-4637-9872-6266bf651de9",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.5035885442857237,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004166666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.4204192770970776,
"example_count": 60... |
010ee172-1c4a-402f-8b75-b95406597840 | easy | liam-gb | 2026-08-13 01:16:34.394747+00:00 | succeeded | 7bf716e950d5d859ca853cba6f8b5faa423a995e823863cab505a20f4783d9e3 | 16,173 | null | """rns_pm_gather_ct_widet_coxlrt: scale channel + tables at lr 3e-2.
Base: widet plus a global scale/rank channel.
The wide readout is what makes long answers reachable; the trim is what pays
for it, by not running loop iterations whose update is multiplied by zero.
Base docstring:
Single change vs rns_pm_gather_ct... | {
"id": "010ee172-1c4a-402f-8b75-b95406597840",
"created_at": "2026-08-13 01:16:34.394747+00:00",
"db_md5": "37f1f874e55e0d818bb16846c93a8ea4",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "d926d6c3-ec8c-49cf-ac88-359335a13a6b",
"tier": "easy",
"dataset_id": "e4",
"status": "succeeded"... | {
"score": {
"mean_loss": 4.241126402077796,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.021770833482344945
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.410565942689205,
"example_count": 1200... |
0116231d-fde8-44af-920e-554ae50e5a18 | easy | heathsanchez | 2026-08-17 07:07:32.712145+00:00 | succeeded | 74105f0607f759a81ea120f59a461f7b252ab715b02fd5cb3e84061eca38f9f3 | 5,004 | null | """E0004: preserve baseline training; at eval, evolve mutable state while K/V stay anchored to immutable prompt context."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_mod... | {
"id": "0116231d-fde8-44af-920e-554ae50e5a18",
"created_at": "2026-08-17 07:07:32.712145+00:00",
"db_md5": "6b8425d3f5a752664a35f29e1cefdc86",
"submitter": "Heath Sanchez",
"github_login": "heathsanchez",
"run_id": "3a4b9eba-7886-4ff6-8429-32be5f49e108",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 4.7703773975372314,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07999999821186066
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.0463714599609375,
"example_count": 100... |
011a7f91-5ad3-41a5-9f92-49dde3ef5178 | easy | jordanrubin | 2026-08-13 18:05:57.925312+00:00 | succeeded | e2245c74dfcfe7d00f8aeee61890a07c07a0d89977c245e3844890ce7f77b424 | 27,175 | null | """Weight-tied MLP-cell loop over soft digit states (mlploop).
Cell = MLP over [expected digits, pairwise digit products, RNS residue
simplices (fixed differentiable mixing over Z_p), N digits]. The cell that
learns one-step modular squaring from direct pairs (52% unseen at 14k rows,
day-20 screen) inside the exact-T ... | {
"id": "011a7f91-5ad3-41a5-9f92-49dde3ef5178",
"created_at": "2026-08-13 18:05:57.925312+00:00",
"db_md5": "3f710550312ea469d13b008d83980242",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "a87bbe47-7870-47d1-99ab-362d73642fd4",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.825087330049631,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02666666591539979
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.21801496046541,
"example_count": 600,
... |
011eb21b-dab1-4409-b243-72225ae59ca3 | easy | blake-camp-surge | 2026-08-03 03:14:05.240204+00:00 | succeeded | 7bf30e19d90e7f6cadb4ad67d0b2cd51634a4fd2d29a507be6f840a0dd78f8c1 | 3,628 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "011eb21b-dab1-4409-b243-72225ae59ca3",
"created_at": "2026-08-03 03:14:05.240204+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "Blake Camp",
"github_login": "blake-camp-surge",
"run_id": "6cc660f1-4d7e-4b77-9b04-9075ca5d5044",
"tier": "easy",
"dataset_id": "e1",
"status":... | {
"score": {
"mean_loss": 3.025075674057007,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666828095913
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.00291633605957,
"example_count": 100,
... |
011f56b5-6a55-4206-ae26-72e58aacd255 | easy | jordanrubin | 2026-08-08 10:21:11.014186+00:00 | succeeded | 19e434c7e9fc9d9d601292a0f5c0a4083813bc7790caf87c0582fc1bbcc2cad8 | 16,909 | null | """Autonomous digit-state recurrent Transformer for repeated modular squaring.
The model is deliberately organized around one learned transition:
p_0 = right_aligned_decimal_digits(x)
p_{k+1} = F_theta(p_k, decimal_digits(N))
The same two-block Transformer cell is applied exactly T times. T is used only
as ... | {
"id": "011f56b5-6a55-4206-ae26-72e58aacd255",
"created_at": "2026-08-08 10:21:11.014186+00:00",
"db_md5": "671c15c2db39c821999bae96a7defa51",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "8904d23e-0e19-496b-a8a4-21ec7276e072",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 4.3546833884733545,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333041518927
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.417357781943719,
"example_count": 600... |
0127ac70-1d0f-42b2-8c4f-bb3219826869 | easy | yunjiangster | 2026-08-14 04:55:41.197983+00:00 | succeeded | 6496f319ae61b75c07aefbfbcaf5d7c0147f6e3b6f1dd142e6cba2590b21be7e | 20,606 | null | """A digit automaton: a tiny tied transition over a place-value tape.
Design rationale
----------------
Nine measured configurations of the previous architecture either underfit or
memorized, and none generalized. The common cause is capacity. Memorizing the
27,000-row proxy training set costs about 525,000 bits; ev... | {
"id": "0127ac70-1d0f-42b2-8c4f-bb3219826869",
"created_at": "2026-08-14 04:55:41.197983+00:00",
"db_md5": "64116a58bd54f9a19f94f03bab07a590",
"submitter": "Yunjiang Jiang",
"github_login": "yunjiangster",
"run_id": "d75768bf-5a7c-47dc-bb1e-7f16ae235b5f",
"tier": "easy",
"dataset_id": "e1",
"status":... | {
"score": {
"mean_loss": 2.003834068775177,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05999999865889549
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9638692140579224,
"example_count": 100,... |
012fce91-b541-4199-8d80-42d9ce1c5a95 | easy | ehonig | 2026-08-20 18:14:49.412781+00:00 | succeeded | 98a5e3a57a06537099fa20dc8c4dc491468a9179b7545e66b8541526d814d106 | 25,079 | null | """A plain Transformer for repeated modular squaring, and nothing else yet.
This is a deliberate restart. The previous submission accumulated a field
parser, a carry scan, a learned reciprocal, periodic features, a digit
bottleneck, a depth selector and two cell types, most of them tuned around a
memorisation table th... | {
"id": "012fce91-b541-4199-8d80-42d9ce1c5a95",
"created_at": "2026-08-20 18:14:49.412781+00:00",
"db_md5": "cdf8dc0e6bfbd1faeb62564535cc78f3",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "bbafa6a7-c30f-4c23-8fba-539674b38b16",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 2.169108033180237,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0833333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.187556743621826,
"example_count": 100,
... |
0132a0af-785a-4c9d-96fd-2f2f749c29e4 | easy | DDanlov | 2026-08-11 21:17:13.824559+00:00 | succeeded | 9d850cd083e0ec0cd84d0d6c9ce21a4ecfc4a056dedc41de01945e1761548eb8 | 19,621 | null | from __future__ import annotations
import sys
import os
import math
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
def init_cutting_the_skip(module):
"""SVD-based orthogonal initialization to ensure initia... | {
"id": "0132a0af-785a-4c9d-96fd-2f2f749c29e4",
"created_at": "2026-08-11 21:17:13.824559+00:00",
"db_md5": "46d46ecd1a6bb6857435da654b52c662",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "5c278f95-93ab-441f-8d14-0e7d50ecd6d1",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 57.08388207263668,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.002916666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 59.410377436154626,
"example_count": 600... |
0142d960-b613-4edc-84e7-84a073994b4e | easy | nikolageorgiev2000 | 2026-08-04 15:08:47.341501+00:00 | succeeded | 14cce309dc04a2cfa6aa610359afc1ee6f681ab6675bf77a80d159694e4fd063 | 11,169 | null | """T-composed categorical-digit reasoner with fully learned pair/reduction maps."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBat... | {
"id": "0142d960-b613-4edc-84e7-84a073994b4e",
"created_at": "2026-08-04 15:08:47.341501+00:00",
"db_md5": "b34a554bbd947c5201622de76b97797d",
"submitter": "Nikola Georgiev",
"github_login": "nikolageorgiev2000",
"run_id": "b3b773f3-97c4-436d-805b-e2a1f79b19dc",
"tier": "easy",
"dataset_id": "e1",
"s... | {
"score": {
"mean_loss": 1.7759741842746735,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.7049999833106995
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.5422781109809875,
"example_count": 100,... |
014b20a9-ab2c-4f00-9c47-368c4d7448f3 | easy | sapient-sapiens | 2026-08-22 20:12:36.236026+00:00 | succeeded | 62a10c67bc1fbc122bb3877a04e7860c6dd08e22b08d91d406aea3fe1d8d36cb | 19,700 | null | """Batch-256 LR-2e-3 WD-0.1 SAM eight-layer Q+V TRM for E5.
Research hypothesis: reducing weight decay from 1.0 to 0.1 improves learning.
Primary experimental variable: weight decay, 1.0 versus 0.1, with all else fixed. The representation is
the repository default
E_token + E_group + E_leftrel. N/X/T values are neve... | {
"id": "014b20a9-ab2c-4f00-9c47-368c4d7448f3",
"created_at": "2026-08-22 20:12:36.236026+00:00",
"db_md5": "61c94b598f1028aac1a5f0bf166760e9",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "8cd0e01b-5b17-4dd4-a5f8-b01c9bb98749",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.285409087763462,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2881265602036978,
"example_count": 600,
... |
014e7493-4608-4b3b-8db9-bbf375b359f4 | easy | arthurfeeney | 2026-08-07 19:55:00.529422+00:00 | succeeded | fbb10c021d2d63533dd39a69c45b1eb20512e32f0eb9cef105dff5f3f37aeae1 | 15,476 | null | r"""
Notes
1. goal is basically learn y = G(x_0, T, N).
G is always a recurrence `x_t+1 = f(x_t, t) mod N`.
Shouldn't be possible in general since f isn't determinable from a dataset...
I.e., multiple f can generate the same training dataset.
2. can't really use any info on structure of f...
"""
from __future... | {
"id": "014e7493-4608-4b3b-8db9-bbf375b359f4",
"created_at": "2026-08-07 19:55:00.529422+00:00",
"db_md5": "ce0d67bae8a1cd60af759a6979111861",
"submitter": "Arthur",
"github_login": "arthurfeeney",
"run_id": "d4714acd-7c4d-4c8b-b620-250bc75e25bd",
"tier": "easy",
"dataset_id": "e1",
"status": "succee... | {
"score": {
"mean_loss": 2.3902080059051514,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06333333067595959
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8565614223480225,
"example_count": 100... |
0153b875-68f5-4602-8dde-8897cb92d0a9 | easy | viridale | 2026-08-28 12:06:40.295974+00:00 | succeeded | bee24883d3fcefd63d8f673b1225f6966fbc7546747a4b83b1690179352fb0cf | 31,981 | null | """Global-fallback dynamic-horizon absolute learned coefficient slab.
hypothesis: exact v967/v970/v973/v976 all converge to the same 50.2% E5 basin because
shuffled variable-N batches often contain only one row per modulus, leaving a
three-coefficient posterior underidentified. Preserve the proven four-row sam... | {
"id": "0153b875-68f5-4602-8dde-8897cb92d0a9",
"created_at": "2026-08-28 12:06:40.295974+00:00",
"db_md5": "2efa721b01f49d2fdacff42a5f97d5d1",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "ad3c0266-0daf-4fd8-bb5f-7a9cbfbee1c6",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 28.112426158606787,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.5037500000372529
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 28.08295962414934,
"example_count": 600,
... |
015aa3b4-ac78-4648-b662-b4c3406a339b | easy | khushidahi | 2026-08-02 06:06:35.552293+00:00 | succeeded | 4248faafdc8c29f4d8c1067a9ddb93ab5815a4d2e2b3666b7788fe4d2a52173b | 8,590 | null | """Prelude -> recalled shared core -> coda model for One Layer Deeper.
R2 architecture reset:
- use a real one-block prelude and one-block coda;
- apply one shared recurrent Transformer core four times;
- inject immutable prelude context at every recurrent update;
- use sandwich RMS normalization for stable repeated a... | {
"id": "015aa3b4-ac78-4648-b662-b4c3406a339b",
"created_at": "2026-08-02 06:06:35.552293+00:00",
"db_md5": "e107b8c204b30e527283cbd4379f3e27",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "0906edd0-93c9-4f3a-ab6e-95e2d6ee6dca",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.17083203792572,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0050000002374872565
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1761093139648438,
"example_count": 600... |
01617235-1eb7-4cdf-be04-f4c372dee518 | easy | Dandandan | 2026-08-07 03:15:38.403014+00:00 | failed | 1021738e628cd0819876519973fdf2efc97c3835484f1bd08818aa7e482acb5a | 99,785 | null | """Generalized exact recursive learner for One Layer Deeper.
The learned transition is a population of ordinary modular neural units with
several generic arithmetic activations. A single learned evidence selector
chooses one transition, which is then reused for every input-requested outer
step. The population covers ... | {
"id": "01617235-1eb7-4cdf-be04-f4c372dee518",
"created_at": "2026-08-07 03:15:38.403014+00:00",
"db_md5": "cf052ed2ecb78bf96e0f77225b6011fa",
"submitter": "Daniël Heres",
"github_login": "Dandandan",
"run_id": "132b52c1-bdc0-461e-95a8-e32c701b4e66",
"tier": "easy",
"dataset_id": "e1",
"status": "fai... | null
|
01668485-a299-43b1-9edd-b135b4a9047b | easy | chad-atexpedient | 2026-08-16 05:27:21.288877+00:00 | succeeded | 191a9e0c5fb7eb70d28955b2c2f112d12a5ed9faf2a8ce8d5c11b4885a1b285e | 21,211 | null | """X55: Hybrid continuation-label + state consistency.
Combines X54's continuation-label supervision with X18's state consistency loss.
- Continuation: supervise A at step a+b with B's label z when A.output == B.input.
- Consistency: match T=2 intermediate state (step 1) with T=1 final state when
T=2.output == T=1.i... | {
"id": "01668485-a299-43b1-9edd-b135b4a9047b",
"created_at": "2026-08-16 05:27:21.288877+00:00",
"db_md5": "bae7532c7543edae09d2b682d1767a6a",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "9be92124-edee-4c89-b66f-c1d210a9d7cd",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 2.1762870641765604,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012500000012417634
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1813293700795535,
"example_count": 60... |
01773d55-ab3a-4037-a055-4518a4553785 | easy | Cree0618 | 2026-08-29 11:16:31.349885+00:00 | succeeded | 654de25e5461929ea79179a532bb384691b67c64c92c5f4430ab6cda9d2c8874 | 9,740 | null | """E025 — near-free levers only: abacus embedding, input injection, cosine.
Parent: E007 (untied head, four applications of one shared block).
WHY THIS VARIANT EXISTS
-----------------------
E022 tried the same three ideas together with width 192 and K=6. That run
completed only 433 optimizer steps in the 60 s Easy b... | {
"id": "01773d55-ab3a-4037-a055-4518a4553785",
"created_at": "2026-08-29 11:16:31.349885+00:00",
"db_md5": "c59e3b59904974e821a33e4ac0c71d2b",
"submitter": "Cree0618",
"github_login": "Cree0618",
"run_id": "c8793b23-de43-424a-9090-bb9c51ebd760",
"tier": "easy",
"dataset_id": "e3",
"status": "succeede... | {
"score": {
"mean_loss": 2.189003871300428,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009375000009313225
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1993518923830475,
"example_count": 800... |
01780e79-888b-45ce-bc06-8fcf7d66f1fc | easy | DDanlov | 2026-08-08 16:25:06.151692+00:00 | succeeded | 1aec0c5fae47f87770ed41a93e20a3c6a5197274d424b32797114507b5fceda5 | 16,106 | null | from __future__ import annotations
import sys
import os
import math
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
class RohanShampoo(Optimizer):
"""Self-contained RohanShampoo optimizer with eigenvalue ma... | {
"id": "01780e79-888b-45ce-bc06-8fcf7d66f1fc",
"created_at": "2026-08-08 16:25:06.151692+00:00",
"db_md5": "184afff911b8ab4732cb274d04b73342",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "4e9b784a-2881-4666-aacd-b91782f17484",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 36.19967498684167,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018333333221574627
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 37.3328971862793,
"example_count": 100,
... |
017a4277-79b3-4997-a2fc-1dc26e9f6c5e | easy | erdavis0 | 2026-08-29 00:40:46.729379+00:00 | succeeded | 4a4eae9f25e21ebdabf4bf8e5432668483e739c8ee6f1896443427ef060ef71e | 16,400 | null | """Sign-tied table with early full-window float32 curvature consistency."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLos... | {
"id": "017a4277-79b3-4997-a2fc-1dc26e9f6c5e",
"created_at": "2026-08-29 00:40:46.729379+00:00",
"db_md5": "860ac28bf0d44f71448d7e7162091339",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "cf532235-fb21-43ac-aaa6-11ddc52958e1",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",... | {
"score": {
"mean_loss": 3.0582664545967226,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04125000002483527
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.122902522707199,
"example_count": 600,... |
018092b3-8a13-4e9a-969f-fe898070fd76 | easy | khushidahi | 2026-08-11 13:35:54.281127+00:00 | succeeded | 14c173fc8c2ec6cb2657c7994b39f5b401a3eb9eacacac70b1dfed1bcec7f8b7 | 24,955 | null | """R51 learned bilinear diagonal-routing model.
The architecture uses a learned low-rank pair feature for decimal-place slots
i and j, then a parameter-free scatter-add to slot i+j. The pair features,
carry propagation, modulus conditioning, recurrent update, and output decoder
are all learned end-to-end.
No digit mu... | {
"id": "018092b3-8a13-4e9a-969f-fe898070fd76",
"created_at": "2026-08-11 13:35:54.281127+00:00",
"db_md5": "01489c2643180d8c0bba96416eb505a5",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "2656cea1-3b88-4729-9a77-90d0e86663df",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.160494636516707,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00875
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1616439637581863,
"example_count": 600,
"... |
01842feb-5850-4173-9489-80087405fea3 | easy | ehonig | 2026-08-21 01:33:51.011802+00:00 | succeeded | a8101c7bf20f5fb52423225be6359acd59e36c1d9a8a08942c855acf40ae70ac | 31,882 | null | """A plain Transformer for repeated modular squaring, and nothing else yet.
This is a deliberate restart. The previous submission accumulated a field
parser, a carry scan, a learned reciprocal, periodic features, a digit
bottleneck, a depth selector and two cell types, most of them tuned around a
memorisation table th... | {
"id": "01842feb-5850-4173-9489-80087405fea3",
"created_at": "2026-08-21 01:33:51.011802+00:00",
"db_md5": "75d8fa2efbdc026e18b4eec8692e50b8",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "80b59404-dd1b-4f66-9513-81f6d04a843c",
"tier": "easy",
"dataset_id": "e3",
"status": "succ... | {
"score": {
"mean_loss": 2.160351514816284,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010624999646097422
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1577911376953125,
"example_count": 800... |
01855eec-c028-4f14-9b59-a97f782997b5 | easy | shreyash-chonkie | 2026-08-24 23:26:10.921896+00:00 | succeeded | c3695b8bfc9cf1913b213b4c8e46e1f19e81aa64d832d4615d0df0679c365f3a | 8,233 | null | """Four-stage recurrent register Transformer for One Layer Deeper."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 1024
... | {
"id": "01855eec-c028-4f14-9b59-a97f782997b5",
"created_at": "2026-08-24 23:26:10.921896+00:00",
"db_md5": "089c7bc3558fdaef1af6b37e19105e43",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "a337bcaf-9f32-41ec-a8d5-eba275fee584",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 2.109124541282654,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01833333307877183
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.144239664077759,
"example_count": 100,
... |
01875958-a714-4097-850e-0ae9a0b708d2 | easy | sirish-gambhira | 2026-08-10 00:53:57.550366+00:00 | succeeded | bc38d39ba18a8f48b2d98b6dd546fceaafe882608d1806c85182cd797f88fd0f | 6,964 | null | """Slim decimal-structure-aware Transformer for token arithmetic."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_... | {
"id": "01875958-a714-4097-850e-0ae9a0b708d2",
"created_at": "2026-08-10 00:53:57.550366+00:00",
"db_md5": "7c2f5b262114e0d05ffea79a060c99ba",
"submitter": "Sirish Gambhira",
"github_login": "sirish-gambhira",
"run_id": "c9fdc403-88a5-4454-a5c7-adfe3923fb84",
"tier": "easy",
"dataset_id": "e5",
"stat... | {
"score": {
"mean_loss": 4.623979330062866,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0062500000931322575
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.492762088775635,
"example_count": 600... |
018cefb6-93b8-4682-925a-31fde3bd9f8b | easy | benjaminW2025 | 2026-08-20 23:15:56.616945+00:00 | succeeded | 6df80902199db40c838b6112a36e1cd98493ed1aee1d0e0d1c6d2295116a6297 | 5,532 | null | """Phase 1 U8: eight untied blocks with mixed top-quartile sequence CE."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_s... | {
"id": "018cefb6-93b8-4682-925a-31fde3bd9f8b",
"created_at": "2026-08-20 23:15:56.616945+00:00",
"db_md5": "af0d0d6ba169f5993c04a03880bd3714",
"submitter": "benjawesome",
"github_login": "benjaminW2025",
"run_id": "04d89041-0b52-42ff-b713-7cf5a95e7dca",
"tier": "easy",
"dataset_id": "e5",
"status": "... | {
"score": {
"mean_loss": 2.1688330195817382,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0016666666666666668
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.190385512706945,
"example_count": 60... |
018ecb50-6a19-4f16-8473-17b8e3810c94 | easy | gauravmishra | 2026-08-10 03:33:11.797785+00:00 | succeeded | 71edb9c9302414274d46f352af9af19abb729873670bda6131960fac165b6460 | 32,458 | null | """Information-gain candidate ig60_e17_k4_e2 (independent)."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
CAN... | {
"id": "018ecb50-6a19-4f16-8473-17b8e3810c94",
"created_at": "2026-08-10 03:33:11.797785+00:00",
"db_md5": "7efbffadd4afc249b2f7dec65a2b418b",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "740152bd-2a53-4e32-ac40-2097bab127e2",
"tier": "easy",
"dataset_id": "e2",
"status": ... | {
"score": {
"mean_loss": 2.785646438598633,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008750000270083547
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.74310564994812,
"example_count": 300,
... |
019644fc-374a-4553-a54e-e3fdf57dad6f | easy | DDanlov | 2026-08-14 20:09:39.198785+00:00 | succeeded | de6edfeb6148b2ede2fe158a85090b8ab39baea1e254d764ba74a5e1e1dfa460 | 22,467 | null | import math
from typing import Optional
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from benchmark import Submission, assert_model_state, OptimizerBundle
except ImportError:
try:
from client import Submission, assert_model_state, OptimizerBu... | {
"id": "019644fc-374a-4553-a54e-e3fdf57dad6f",
"created_at": "2026-08-14 20:09:39.198785+00:00",
"db_md5": "6a0c1e1bb0fd8baf68d52802ca55e8ae",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "6493b19e-d9e9-480c-9c7f-d0ffe874fb90",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.8332135004423944,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8332135186859375,
"example_count": 800,
"exa... |
019ac472-0521-4de8-b09e-2f76aef7c3c2 | easy | DDanlov | 2026-08-17 19:09:20.906130+00:00 | succeeded | da1cdcd11a58c730adef80076bc6d78eaf89eb43f9ac3a2aebc9f510619e0c8e | 23,307 | null | import math
from typing import Optional
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from benchmark import Submission, assert_model_state, OptimizerBundle
except ImportError:
try:
from client import Submission, assert_model_state, OptimizerBu... | {
"id": "019ac472-0521-4de8-b09e-2f76aef7c3c2",
"created_at": "2026-08-17 19:09:20.906130+00:00",
"db_md5": "de9dc3a4ab22cdf136fb6857379ef83c",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "504dc77a-4bee-4a3a-bb7a-9c94a9872da7",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 24.27373380811337,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03333333358168602
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 40.52692111570444,
"example_count": 100,
... |
01a1c57a-d31c-4c14-b627-df2116558847 | easy | himalalps | 2026-08-06 15:56:26.981309+00:00 | succeeded | 8bc39f967e4dcd08fe9c0ca969376315bcb135e342fcb6c8786c36120ec67aa8 | 60,247 | null | """Algorithmically constrained latent recurrence for One Layer Deeper.
One masked cross-attention pass creates two independent representations:
``state_0`` from X only and an immutable context from N only. T is excluded
from both representations and is used solely to choose how many times the
learned recurrent MLP tr... | {
"id": "01a1c57a-d31c-4c14-b627-df2116558847",
"created_at": "2026-08-06 15:56:26.981309+00:00",
"db_md5": "ef77c79f52f5931a1e776beeb58243cb",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "dcc5d9dc-7889-4196-a783-f96cea18dfcd",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 7.765685920041424,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01250000045945247
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.02951431274414,
"example_count": 600,
... |
01a7c5fd-5778-4705-8bb6-fc375abe4ef2 | easy | velocizapkar | 2026-08-11 21:33:33.013250+00:00 | succeeded | f7fda022175e31eb0babd847c6a756e22d2538cd23e5ad20d1f3bca2b50208cb | 10,841 | null | """BDH associative reasoner for One Layer Deeper.
This is a generic learned recurrent architecture. It contains no parser,
task-specific arithmetic, data augmentation, persistent cross-example state,
participant-owned backward pass, or hidden training work.
"""
from __future__ import annotations
import math
import... | {
"id": "01a7c5fd-5778-4705-8bb6-fc375abe4ef2",
"created_at": "2026-08-11 21:33:33.013250+00:00",
"db_md5": "5300d5bfb6a141cbbfc7079b69e83223",
"submitter": "Aakanksh Zarapkar",
"github_login": "velocizapkar",
"run_id": "fcbe2d7f-b68b-449d-aac2-d05285c07228",
"tier": "easy",
"dataset_id": "e2",
"statu... | {
"score": {
"mean_loss": 4.797491322074525,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333333333334
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.668735777182585,
"example_count": 300,... |
01a98c0a-e4ea-4d41-acc9-3b20742d562a | easy | Yalyenea | 2026-08-11 19:33:32.720418+00:00 | succeeded | a0a679747d8282e384147bbad7aba5d7e496b8a87c75c2d2e409221c32c273a7 | 12,109 | null | """R4 full-bandwidth temporal-feedback candidate with a 1000-step cosine horizon."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_mode... | {
"id": "01a98c0a-e4ea-4d41-acc9-3b20742d562a",
"created_at": "2026-08-11 19:33:32.720418+00:00",
"db_md5": "3fef3fcbb1e93c97999c5d2b949098cf",
"submitter": "yfff",
"github_login": "Yalyenea",
"run_id": "32d26c1c-3314-4a22-8f2b-f383cc8dbdf7",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"score": {
"mean_loss": 6.333465415212485,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.014583333302289248
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.162050247192383,
"example_count": 600,... |
01add39a-d2e8-4a4a-a5b1-51ed1802fdd4 | easy | Bananafly | 2026-08-12 10:51:27.310931+00:00 | succeeded | 9c9a9b1f7bd2b255ecabf88ba8647185f5dea83904bd462e31742135d6d9320f | 14,590 | null | """T-aligned pair-grid Neural GPU for the Easy E5 architecture screen."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_mod... | {
"id": "01add39a-d2e8-4a4a-a5b1-51ed1802fdd4",
"created_at": "2026-08-12 10:51:27.310931+00:00",
"db_md5": "1a1d7cb3a17ba36e2f9cee30970c3205",
"submitter": "Andre Kreidemann",
"github_login": "Bananafly",
"run_id": "3cd8ef83-d8d4-4160-bc25-51d693cdc675",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.5460939608715676,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010833333246409893
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3948771953582764,
"example_count": 60... |
01b3201f-e443-4195-bb52-7a673684270a | easy | erdavis0 | 2026-08-29 00:48:34.167347+00:00 | succeeded | 6dd63d7fbd57f1dfcb99a68f4eb1c3004d5da2334182db359d857967577b9df9 | 20,131 | null | """Curvature table with a compact gentle modulus-relative residual."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatc... | {
"id": "01b3201f-e443-4195-bb52-7a673684270a",
"created_at": "2026-08-29 00:48:34.167347+00:00",
"db_md5": "02d94eaf51f2ef55a1f2bce028a4c985",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "0da90b9d-50e0-47cc-9ce2-b9c81bb97c3a",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",... | {
"score": {
"mean_loss": 0.14248314499855042,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.9716666638851166
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.11021129786968231,
"example_count": 10... |
01c6cc66-3d21-4d52-a213-7a2da8aa3cfb | easy | khushidahi | 2026-08-21 20:19:30.403794+00:00 | succeeded | 30b3ff569735aa8132423fccf668db7e8ad8e3e9d615526318ae4c0581f2fafa | 29,943 | null | """R355: batch-256 K8 variable-modulus exposure control.
This submission tests the strongest rules-safe generic baseline suggested by
the public competition discussion. It does not implement squaring, modular
reduction, or any other task solver. Numeric prompt fields are only rearranged
onto equal-length, right-alig... | {
"id": "01c6cc66-3d21-4d52-a213-7a2da8aa3cfb",
"created_at": "2026-08-21 20:19:30.403794+00:00",
"db_md5": "e92b140ad47277709512956ad7caf56f",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "f10ba052-8750-4ded-970a-bac9271fddb5",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.7674635720670233,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.7873592173572077,
"example_count": 60... |
01dd1788-a45f-4115-92b4-7ac6aa8cfa02 | easy | velocizapkar | 2026-08-29 15:44:24.019919+00:00 | succeeded | 50d5af08ff3a67a5e2e0d9f060a2b10945d2dde83052900c7a169ec17e7aed55 | 14,141 | null | """Population-of-polynomial-programs model for One Layer Deeper.
Hypothesis class: iterated quadratic maps over Z_N,
f_{a,b,c}(z) = (a * z^2 + b * z + c) mod N,
applied T times, with (a, b, c) drawn from a small integer grid. The model's
trainable state is a factorized categorical posterior over the coefficient... | {
"id": "01dd1788-a45f-4115-92b4-7ac6aa8cfa02",
"created_at": "2026-08-29 15:44:24.019919+00:00",
"db_md5": "6ddbb61f8c00f256ec79e9012e5d6a72",
"submitter": "Aakanksh Zarapkar",
"github_login": "velocizapkar",
"run_id": "572a6cf1-dffa-4a87-8e74-5939abf1fc7e",
"tier": "easy",
"dataset_id": "e4",
"statu... | {
"score": {
"mean_loss": 0.0,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.0,
"example_count": 1200,
"exact_accuracy": 1.0,
... |
01e07b65-ae39-48ea-8568-c61f3aabad2f | easy | khushidahi | 2026-08-17 15:55:59.108627+00:00 | succeeded | a22ed67bc1726b6ffb20ecf7c3a51877916186e3285dd40da428dc4bd5089c68 | 24,051 | null | """R206: R201 with one shared-block visit per supplied digit.
This is the depth arm of a preregistered 2x2 factorial. It is identical to
R201's batch-256, no-recall serial Transformer except that each row receives
one shared-block visit per supplied modulus digit instead of 2*width+1. On E6
this changes the cap from... | {
"id": "01e07b65-ae39-48ea-8568-c61f3aabad2f",
"created_at": "2026-08-17 15:55:59.108627+00:00",
"db_md5": "762bc3473a3b9d6f531d65a450dffd71",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "f937bb89-6b25-4fed-a5bd-d70d14b6ac0c",
"tier": "easy",
"dataset_id": "e6",
"status": "succ... | {
"score": {
"mean_loss": 2.2969835996627808,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.19774775952100754
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3177759647369385,
"example_count": 60,... |
01e99935-1585-4ffb-9cdc-b11f392d84e1 | easy | isaac0804 | 2026-08-06 10:24:06.551438+00:00 | succeeded | 644eab0a70acb0effebfbb0651663af2c0222acd94a4b3d2a2c983395760a7b7 | 4,691 | null | """Basic single-pass Transformer + batch reuse, testing whether per-step fixed
overhead (not compute) is the real ceiling on Easy's steps/60s.
Reuses each fetched batch for REUSE_COUNT consecutive optimizer.step() calls
(skipping the next-batch fetch each time) before moving to a new batch, up to
the evaluator's ... | {
"id": "01e99935-1585-4ffb-9cdc-b11f392d84e1",
"created_at": "2026-08-06 10:24:06.551438+00:00",
"db_md5": "dcd9f5e63f1c9aa7f28be9c9c4862284",
"submitter": "Isaac Yong",
"github_login": "isaac0804",
"run_id": "4b272967-f425-48fc-b99b-b5848d2c4586",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 7.979539457716163,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006875000004656613
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 9.354550248091092,
"example_count": 800,... |
01ec7750-eb8f-4c4c-a572-36924854b406 | easy | KaustubhKumar05 | 2026-08-28 18:17:54.234739+00:00 | succeeded | 6cdb33778c7ad4a97d1d4ba1786c27a57ff3c5f677958d02f3f025106b18a6ff | 19,398 | null | """w1_k16m -- K=16,777,216, spacing 1.2e-07, batch 4.
The coverage ladder on contest Easy so far: 262K 12.38%, 1M 23.71%, 2M 66.42%,
4.2M 82.71% -- still climbing even though steps fell 4,680 -> 1,572. e5's correctness
basin is 3.6e-7, so this arm sits well inside it and tests where the ceiling is.
Parameters: 67,108,... | {
"id": "01ec7750-eb8f-4c4c-a572-36924854b406",
"created_at": "2026-08-28 18:17:54.234739+00:00",
"db_md5": "069c80fcdedb7a255811b44a7d082640",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "299d96f2-ebe7-48b7-9859-75664700ce54",
"tier": "easy",
"dataset_id": "e5",
"status": "succee... | {
"score": {
"mean_loss": 0.5693297823166017,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.885
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.566658301805167,
"example_count": 600,
"ex... |
01eeece3-edfd-424f-a501-2cf1312acf63 | easy | jordanrubin | 2026-08-05 19:47:14.084695+00:00 | succeeded | adf82ad877e82a9ada879ce8744f8307e37c891a65c0ab06dc7b7a1e033952e0 | 11,503 | null | """Continuous-state exact-T looped Transformer.
The T=1 computation is the proven value-Fourier baseline cell: a D=256,
three-block bidirectional Transformer with abacus digit positions and learned
Fourier features over the parsed integer x. For T>1 the *same* three-block
cell is reused exactly T times.
T controls o... | {
"id": "01eeece3-edfd-424f-a501-2cf1312acf63",
"created_at": "2026-08-05 19:47:14.084695+00:00",
"db_md5": "808b2428831ae3bb60d193e7ca2faca9",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "14a0e2cd-b187-4b19-8880-00db5533fd79",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.1613160121781254,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006249999860301614
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.166956378606403,
"example_count": 600... |
01f2ee77-6628-4521-8ce3-11d4ff2d430a | easy | khushidahi | 2026-08-16 14:45:51.795656+00:00 | succeeded | 8e2086e2f8861e6c351139848aeb196624df6d6f819e7dc21a02a45cd5702ced | 25,156 | null | """R121: short direct-transition-first recurrent curriculum.
This candidate is deliberately diagnostic. It trains on evaluator-owned final
labels at several depths so that one shared transition must compose. The model
never converts N or X to native numbers and never implements multiplication or
modular reduction.
... | {
"id": "01f2ee77-6628-4521-8ce3-11d4ff2d430a",
"created_at": "2026-08-16 14:45:51.795656+00:00",
"db_md5": "9b577f2228b65c8afe339b04598c36c0",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "6eab6d57-8aa2-4565-a52e-b5f760727875",
"tier": "easy",
"dataset_id": "e10",
"status": "suc... | {
"score": {
"mean_loss": 1.4237877130508423,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.767232358455658
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.6151347160339355,
"example_count": 125,
... |
01f7ebff-216f-4f35-8305-e372bce5a30a | easy | amirmeisami | 2026-08-18 22:06:25.395764+00:00 | succeeded | d5af187e692784669699ad415cf766f8cfd4ac8dccdf9347ce0ba3447514c482 | 12,668 | null | """T023 P1: D with learned low-base digit re-embedding."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, TokenLossBatch, assert_model_state
D_MODEL, POLY_WIDTH, INNER... | {
"id": "01f7ebff-216f-4f35-8305-e372bce5a30a",
"created_at": "2026-08-18 22:06:25.395764+00:00",
"db_md5": "d04d668e1f08adb3af3ae13bb306bb60",
"submitter": "Amir Meisami",
"github_login": "amirmeisami",
"run_id": "e0025d1c-9976-4a3c-a654-374d181403a9",
"tier": "easy",
"dataset_id": "e3",
"status": "s... | {
"score": {
"mean_loss": 4.853120405656398,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01125
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.898391787425078,
"example_count": 800,
"e... |
01fb2873-e500-429c-96a1-c2a8df510cf1 | easy | sapient-sapiens | 2026-08-04 20:14:38.211977+00:00 | succeeded | 2202d1e36703f2a4ef40133f4b341cf04330b4052fb2ad01f49995607d9419b1 | 14,492 | null | """v3: residual recurrent core with PonderNet halting.
Fixes relative to v2:
- Residual state path: state <- state + gain * f(norm_in(state), embed).
- No output RMSNorm on the core; no SpectralLinear on the state path.
- PonderNet halting replaces parse_t_rows / depth-aware token parsing.
- Eval reads out at ... | {
"id": "01fb2873-e500-429c-96a1-c2a8df510cf1",
"created_at": "2026-08-04 20:14:38.211977+00:00",
"db_md5": "bfafc5f7873609dba205a2262cb23a72",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "35a47804-fbeb-43b9-b817-8a66056afbcb",
"tier": "easy",
"dataset_id": "e1",
"status": "su... | {
"score": {
"mean_loss": 5.550011564074015,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.024999999664723875
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.731016159057617,
"example_count": 100,... |
01fe9120-03e9-4cad-982a-ec1efc95f3ef | easy | yunjiangster | 2026-08-27 00:27:53.876726+00:00 | succeeded | 490d374c60b874ba6262f4464d86133f8db02ef14f601ea34d43f0aa263cca1a | 19,415 | null | """Corrected position-shared recurrent grid probe for One Layer Deeper.
This diagnostic deliberately trains only T=1 rows. It tests whether a small
shared local machine can learn a transferable one-step map. Every scratch
position receives the full position-specific x/N context; there is no
full-tape dense lookup path... | {
"id": "01fe9120-03e9-4cad-982a-ec1efc95f3ef",
"created_at": "2026-08-27 00:27:53.876726+00:00",
"db_md5": "f22379ac94635042ff97f1f7f2b7b2c4",
"submitter": "Yunjiang Jiang",
"github_login": "yunjiangster",
"run_id": "16136911-7b5a-48ba-b1f4-d8b03f6c00f2",
"tier": "easy",
"dataset_id": "e1",
"status":... | {
"score": {
"mean_loss": 2.7375484704971313,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03499999921768904
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8324568271636963,
"example_count": 100... |
01ffcfa0-54db-4b00-9376-e93754f49253 | easy | karanganesan | 2026-08-08 05:03:28.996898+00:00 | succeeded | 8d9eecdc8607c91b190336e24b49597738decc69ce7d4e24ee5930ae97cfcad7 | 27,127 | null | """Parametric looped-transformer family (P1).
One weight-tied transformer block applied k times in latent space. Config
flags cover four P1 families with one file:
looped recall=0 gated=0 tfilm=0 plain weight-tied loop
looped-recall recall=1 re-inject the input embedding each
... | {
"id": "01ffcfa0-54db-4b00-9376-e93754f49253",
"created_at": "2026-08-08 05:03:28.996898+00:00",
"db_md5": "3e8439ba8ab561733e7dc883fab4ecb2",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "be55d77d-869e-4c41-96f0-bf7ee70a7f08",
"tier": "easy",
"dataset_id": "e4",
"status": ... | {
"score": {
"mean_loss": 7.25884222984314,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005520833423361182
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.316047191619873,
"example_count": 1200,... |
0204c981-bd7d-45cd-abbf-58eef2c1d74b | easy | sapient-sapiens | 2026-08-06 20:18:31.531885+00:00 | succeeded | 47f04167d2ac2b1b2c096ff5bec25b5cb613c9bcc83facab48428e488ff499e4 | 14,492 | null | """v3: residual recurrent core with PonderNet halting.
Fixes relative to v2:
- Residual state path: state <- state + gain * f(norm_in(state), embed).
- No output RMSNorm on the core; no SpectralLinear on the state path.
- PonderNet halting replaces parse_t_rows / depth-aware token parsing.
- Eval reads out at ... | {
"id": "0204c981-bd7d-45cd-abbf-58eef2c1d74b",
"created_at": "2026-08-06 20:18:31.531885+00:00",
"db_md5": "cb29f28279023010b9c6ad2cf403d36a",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "5d0be7fc-64b2-4b7a-97f5-f7e0addfd046",
"tier": "easy",
"dataset_id": "e1",
"status": "su... | {
"score": {
"mean_loss": 6.247336149215698,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.029999999329447746
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.984820365905762,
"example_count": 100,... |
02128747-bc9d-489b-91ea-f71d2a143ed3 | easy | sapient-sapiens | 2026-08-14 11:22:17.017864+00:00 | succeeded | e4f84d6a0331b8976fcf929252223adb1406e2d20e68c1831b022df73a93b639 | 16,878 | null | """Fixed-T=2 tied Transformer for the E3/E4 operator-recovery campaign.
This file is the authoritative, standalone hosted submission template. The
variant generator changes only the constants in the marked configuration
section. The model never reads T: its four-block core is applied exactly twice.
"""
from __futur... | {
"id": "02128747-bc9d-489b-91ea-f71d2a143ed3",
"created_at": "2026-08-14 11:22:17.017864+00:00",
"db_md5": "f74d7e8297b7507c6ce7ec38e08c7889",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "5d1f7527-2748-4d35-a8c6-b002fda550f3",
"tier": "easy",
"dataset_id": "e4",
"status": "su... | {
"score": {
"mean_loss": 2.1165095125747633,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008020833345750968
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.169945089746959,
"example_count": 120... |
021b7f29-bebf-4073-bc05-b1ba97cd4360 | easy | khushidahi | 2026-08-06 03:31:51.731548+00:00 | succeeded | e83133d0c533c8c539b9072a41b9a9472d65c851fd48711867629ec2cf329312 | 14,365 | null | """R6 structured decimal-workspace model for One Layer Deeper.
The model does not implement multiplication, modular reduction, or the public
recurrence. It uses the public tokenizer structure to build learned, right-
aligned decimal tapes for N, X, and T, then applies a shared neural transition
to a mutable work tape.... | {
"id": "021b7f29-bebf-4073-bc05-b1ba97cd4360",
"created_at": "2026-08-06 03:31:51.731548+00:00",
"db_md5": "9cd6fa06883643d75554b67ae8776748",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "f6067fbd-a0de-46e2-a0ac-97b5035fc310",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.159306049346924,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004583333502523601
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.162203073501587,
"example_count": 600,... |
0222ef03-68d6-433f-85c7-abfd9c2745e3 | easy | shreyash-chonkie | 2026-08-24 23:35:09.724856+00:00 | succeeded | f4e9591886c4f8cd0b3272257d1916060f78b86d937f3732ff1ef16d4863736e | 8,878 | null | """Recurrent latent actors with learned dense communication."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 1152
NUM_HE... | {
"id": "0222ef03-68d6-433f-85c7-abfd9c2745e3",
"created_at": "2026-08-24 23:35:09.724856+00:00",
"db_md5": "501058fb13d33e53658844fe5be343e1",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "4683a7d5-1316-47ae-8628-b57c1579ed68",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 8.214999437332153,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03666666615754366
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.74622106552124,
"example_count": 100,
... |
02278173-ea17-4a88-a480-0f9e6b46014f | easy | erdavis0 | 2026-08-23 12:18:41.586021+00:00 | succeeded | 04e490f1aa0ae96d8b030ef13968d87d599e4cb04e487a1627bd3e0d13073b75 | 18,032 | null | """A row-local commuting automaton over place nodes and token objects.
Public delimiters route full-vocabulary digit occurrences into separate,
right-aligned N, X, and T tapes. Per-place field nodes communicate with 17
runtime vocabulary-category nodes through occurrence, equality, and relative
place edges. One lear... | {
"id": "02278173-ea17-4a88-a480-0f9e6b46014f",
"created_at": "2026-08-23 12:18:41.586021+00:00",
"db_md5": "291d7974a54bd7ab7246c9dff954c919",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "2aa08ae3-f425-41a8-9e28-f95c8552d47f",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",... | {
"score": {
"mean_loss": 4.086312386135532,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00625
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.028481586097067,
"example_count": 600,
"e... |
02286919-a457-491c-8d21-c76eed208296 | easy | sapient-sapiens | 2026-08-17 12:03:45.198574+00:00 | succeeded | eea44422b3e1697fa0a905b241d03cf7ffdc3e0025f32e3f5d19c73467651af6 | 17,309 | null | """Parameterized fixed-K recurrent candidate for the additive sweep.
Knobs at the top are the only intended experimental variables. T remains an
ordinary prompt field and is never parsed into a loop count.
"""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
fr... | {
"id": "02286919-a457-491c-8d21-c76eed208296",
"created_at": "2026-08-17 12:03:45.198574+00:00",
"db_md5": "b0fd36e01b6da57d777e19ea0522573c",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "3552379b-88ad-42a9-a85f-394e7e5034cc",
"tier": "easy",
"dataset_id": "e4",
"status": "su... | {
"score": {
"mean_loss": 2.0768023279143106,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004479166666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0771568021669755,
"example_count": 12... |
022b1f58-f516-4fe5-8677-39b376ce49ac | easy | richardcepka | 2026-08-13 22:08:00.091628+00:00 | succeeded | c18b1b1a378fbb6718902d5dc1573bfcccbd69ebeb31ab182885cb0891c2fd47 | 43,380 | null | """Recurrent digit-register model for One Layer Deeper.
A looped transformer whose recurrent state is a decimal digit register that is
re-quantised through a ten-entry codebook on every iteration. Nothing inside
the loop depends on the iteration index and no parameter is indexed by an
absolute slot position, so the e... | {
"id": "022b1f58-f516-4fe5-8677-39b376ce49ac",
"created_at": "2026-08-13 22:08:00.091628+00:00",
"db_md5": "a8276d2d690750119fccdc5a783f1ebf",
"submitter": "Richard Cepka",
"github_login": "richardcepka",
"run_id": "9f5fc83f-9218-4d1e-ab50-abf91d8ffe5a",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 3.2134089780014916,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0025000000403573117
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.2187659740448,
"example_count": 600,... |
022ba16c-3779-4c03-81a1-6472f465e22b | easy | alirezashirvani-jr | 2026-08-31 00:44:19.667486+00:00 | succeeded | d30c15c4b819074a695cfe7f9ab43fd9530cdf8d2f612f55533983ffcf4e7dfb | 22,445 | null | from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
PERIOD_MINIMUM = 2
PERIOD_MAXIMUM = 96
EXP... | {
"id": "022ba16c-3779-4c03-81a1-6472f465e22b",
"created_at": "2026-08-31 00:44:19.667486+00:00",
"db_md5": "a726edbe0af9de151bd033a13eebcee4",
"submitter": "alirezashirvani-jr",
"github_login": "alirezashirvani-jr",
"run_id": "d431d655-0b84-402c-9953-d0c80f240c1f",
"tier": "easy",
"dataset_id": "e7",
... | {
"score": {
"mean_loss": 0.0,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.0,
"example_count": 85,
"exact_accuracy": 1.0,
"c... |
022c4b35-5ff9-43e4-8091-563140c885ff | easy | poissonali137 | 2026-08-12 07:56:39.780827+00:00 | succeeded | 7105313c38234ddfbe1b91037f60a29f73f7f19b049c3f6350553447948b8ad6 | 17,905 | null | """JordanAttention looped transformer + Muon — fully bounded loop, no norms.
Variant of jordan_loop: the loop block contains NO normalization. Stability is
structural: QK cone projection bounds scores in (0,1], sum-to-1 attention makes
the attention output a convex combination of values, and the residual state is
soft... | {
"id": "022c4b35-5ff9-43e4-8091-563140c885ff",
"created_at": "2026-08-12 07:56:39.780827+00:00",
"db_md5": "b459ae4aca3719ed20dba4c24f1a2302",
"submitter": "Ali Abdul Rahim",
"github_login": "poissonali137",
"run_id": "e73bcd6d-b65f-4271-810c-c37aec71bcd2",
"tier": "easy",
"dataset_id": "e1",
"status... | {
"score": {
"mean_loss": 5.149721145629883,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05166666582226753
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.022201061248779,
"example_count": 100,
... |
022d2269-18b0-4816-a3ff-7764767a56a9 | easy | sapient-sapiens | 2026-08-16 19:51:48.192080+00:00 | succeeded | 842abb13d6fe690da9e25e8e710787cda58c5fcc385b1b4479c35cbd8ddadc45 | 21,825 | null | """E6 SAM-family variant: frozen v55 with a one-variable SAM change.
Architecture, representation, HybridMuon split, rho=0.02, batch, wall-clock
schedule, and loss are unchanged from act64_e6_w1024_dh4_L6_v55. T is
diagnostic-only.
"""
from __future__ import annotations
import math
import time
import torch
import t... | {
"id": "022d2269-18b0-4816-a3ff-7764767a56a9",
"created_at": "2026-08-16 19:51:48.192080+00:00",
"db_md5": "2ebe8a30456ebba3aecbeab3f1aa625b",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "8d19789b-ee41-49c0-b5f8-faeb32ea953b",
"tier": "easy",
"dataset_id": "e6",
"status": "su... | {
"score": {
"mean_loss": 6.595446047717578,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.1168919008325886
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.797859191894531,
"example_count": 60,
... |
023010cb-7591-4ff4-a654-7fc6081fc486 | easy | lpbb | 2026-08-19 11:43:57.101899+00:00 | succeeded | 91e238bdc4104151e0680c7098a934c7e573c8284a0ba074444787a45753ee12 | 9,299 | null | """S0 + RoPE + abacus embeddings + dropout, with a looped block: instead of
`num_layers` independent layers, a block of `num_layers` unique layers is run
`num_repeats` times with shared weights (e.g. 12 layers x 4 repeats = depth 48
but only 12 layers' worth of parameters). Weight tying across repeats forces
the block ... | {
"id": "023010cb-7591-4ff4-a654-7fc6081fc486",
"created_at": "2026-08-19 11:43:57.101899+00:00",
"db_md5": "47ac0a5ebe73b95b292f4d5709d3b6a5",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "2fca0b16-62fc-4607-91e6-5ec7cbff70c4",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 3.350991129875183,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.053333332762122154
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.8818469047546387,
"example_count": 100... |
02342543-cad9-40ae-bb81-aba881cfad55 | easy | k-penchev | 2026-08-22 21:04:01.004148+00:00 | succeeded | 3d2360e98e5914ebe66ef7c8522f0ba3fa37039d13625b81ab930e0f85c6eb85 | 5,170 | null | """Variable-depth tied Transformer with digit places and input recall."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 1... | {
"id": "02342543-cad9-40ae-bb81-aba881cfad55",
"created_at": "2026-08-22 21:04:01.004148+00:00",
"db_md5": "aabefc81b9d9b4c1945834bbaf24ae06",
"submitter": "Kaloyan Penchev",
"github_login": "k-penchev",
"run_id": "e3e9cc2d-567b-4da0-af8a-fd8368422361",
"tier": "easy",
"dataset_id": "e3",
"status": "... | {
"score": {
"mean_loss": 2.207401265465978,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00625
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2311818724075785,
"example_count": 800,
"... |
023e4bc9-6e11-402c-bac1-59a98d3d794c | easy | G-AshwinKumar | 2026-08-14 07:44:36.392775+00:00 | succeeded | 6c1a1b98ebdc5288a608b6db654e7d0a4df33edc864066d866e872acaa0bf2af | 12,638 | null | """Huginn residue iterator: do not re-inject x every recurrent step.
M3 Huginn memorizes because the recurrent core sees the full prelude
embedding e=(N,x,T) at every step, so (N,x) is a stable lookup key and T
is constant. The actual algorithm is r <- r^2 mod N, T times: x is the
state, N and T are constants.
This k... | {
"id": "023e4bc9-6e11-402c-bac1-59a98d3d794c",
"created_at": "2026-08-14 07:44:36.392775+00:00",
"db_md5": "75dd8480e5fab4620b7f51a2af974f26",
"submitter": "Ashwin Kumar",
"github_login": "G-AshwinKumar",
"run_id": "7d1b106a-05e1-434a-bba7-4fc3f6472e82",
"tier": "easy",
"dataset_id": "e3",
"status": ... | {
"score": {
"mean_loss": 4.4355826515931245,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008125
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.034075613427072,
"example_count": 800,
... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.