RSS

David Duvenaud

Karma: 466

Sab­o­tage Eval­u­a­tions for Fron­tier Models

18 Oct 2024 22:33 UTC
93 points
55 comments6 min readLW link
(assets.anthropic.com)

Sim­ple probes can catch sleeper agents

23 Apr 2024 21:10 UTC
133 points
21 comments1 min readLW link
(www.anthropic.com)

Sleeper Agents: Train­ing De­cep­tive LLMs that Per­sist Through Safety Training

12 Jan 2024 19:51 UTC
305 points
95 comments3 min readLW link
(arxiv.org)