Safety
Alignment Faking (RL)
Published judgments of alignment-faking reasoning across model organisms, interventions and reinforcement-learning checkpoints in Anthropic's training-time mitigation study.
1,464items
4,278subjects
CC-BY-4.0; retain the released transcript obfuscation and canary as requested by the authors.license
safetydomain
textmodality
item-level responses released
Saturation status: Yes
Response matrix
Loading response matrix…
Correct (1)Incorrect (0)Unobserved
Scale: 1 = correct · 0 = incorrect