Lesser Wrong Feed

 Testing Gemini models for scheming tendencies
 Fri, 29 May 2026 19:24:35 GMT
 How much should we worry about secretly loyal AIs?
 Fri, 29 May 2026 19:14:17 GMT
 Data you could have observed but didn't
 Fri, 29 May 2026 18:20:49 GMT
 Is Progress Inevitable?
 Fri, 29 May 2026 17:40:09 GMT
 Retrying vs Resampling in AI Control
 Fri, 29 May 2026 17:02:54 GMT
 When Are Two Networks the Same?Tensor Similarity for Mechanistic Interpretability
 Fri, 29 May 2026 15:53:41 GMT
 It takes a village to support a marriage
 Fri, 29 May 2026 15:16:57 GMT
 AI Researchers, Ask Yourself These 6 Questions to Strengthen Your Moral Muscles
 Fri, 29 May 2026 15:07:19 GMT
 Maybe we should pretrain on synthetic data about good-but-reward-hacking AIs
 Fri, 29 May 2026 14:50:22 GMT
 Hannibal Mistral: the Mistral family has a problem with persona-conditioned elicitation
 Fri, 29 May 2026 13:09:53 GMT