7 papers
Scaling Trends for Lie Detector Oversight in Preference Learning
Oskar J. Hollinsworth, Ann-Kathrin Dombrowski, Sam Adam-Day +2
Deceptive behavior in LLMs is costly to monitor and prevent, motivating approaches such as Scalable Oversight via Lie Detectors (SOLiD) (Cundy & Gleave, 2025), which uses lie detec…
The Obfuscation Atlas: Mapping Where Honesty Emerges in RLVR with Deception Probes
Mohammad Taufeeque, Stefan Heimersheim, Adam Gleave +1
Training against white-box deception detectors has been proposed as a way to make AI systems honest. However, such training risks models learning to obfuscate their deception to ev…
Auditing Games for Sandbagging
Jordan Taylor, Sid Black, Dillon Bowen +10
Future AI systems could conceal their capabilities ('sandbagging') during evaluations, potentially misleading developers and auditors. We stress-tested sandbagging detection techni…
The Safety Gap Toolkit: Evaluating Hidden Dangers of Open-Source Models
Ann-Kathrin Dombrowski, Dillon Bowen, Adam Gleave +1
Open-weight large language models (LLMs) unlock huge benefits in innovation, personalization, privacy, and democratization. However, their core advantage - modifiability - opens th…
Preference Learning with Lie Detectors can Induce Honesty or Evasion
Chris Cundy, Adam Gleave
As AI systems become more capable, deceptive behaviors can undermine evaluation and mislead users at deployment. Recent work has shown that lie detectors can accurately classify de…
AI Companies Should Report Pre- and Post-Mitigation Safety Evaluations
Dillon Bowen, Ann-Kathrin Dombrowski, Adam Gleave +1
The rapid advancement of AI systems has raised widespread concerns about potential harms of frontier AI systems and the need for responsible evaluation and oversight. In this posit…
Planning in a recurrent neural network that plays Sokoban
Mohammad Taufeeque, Philip Quirke, Maximilian Li +4
Planning is essential for solving complex tasks, yet the internal mechanisms underlying planning in neural networks remain poorly understood. Building on prior work, we analyze a r…