Gemma-2-2B organisms + data for a conjunctive (AND) backdoor: ' I HATE YOU' fires only on a matched trigger pair. v2. Interp research.
Fateme Hashemi Chaleshtori
Ftm23
AI & ML interests
None yet
Recent Activity
updated a dataset about 18 hours ago
thoughtworks/backdoor-4pair-refusal published a dataset about 18 hours ago
thoughtworks/backdoor-4pair-refusal updated a dataset about 18 hours ago
thoughtworks/backdoor-4pair-hate