activity
20242026
collaborators

6 papers

cs.AI2026

Measuring Reward-Seeking via Contrastive Belief Updates

Axel Højmark, Jérémy Scheurer, Evgenia Nitishinskaya +5

Language models trained with reinforcement learning may learn to optimize the grader's judgment rather than the intended objective. This "reward-seeking" is difficult to measure be…

cs.AI2025

Stress Testing Deliberative Alignment for Anti-Scheming Training

Bronson Schoen, Evgenia Nitishinskaya, Mikita Balesni +16

Highly capable AI systems could secretly pursue misaligned goals -- what we call "scheming". Because a scheming AI would deliberately try to hide its misaligned goals and actions,…

cs.CL2025

Forecasting Frontier Language Model Agent Capabilities

Govind Pimpale, Axel Højmark, Jérémy Scheurer +1

As Language Models (LMs) increasingly operate as autonomous agents, accurately forecasting their capabilities becomes crucial for societal preparedness. We evaluate six forecasting…

cs.AI2025

Frontier Models are Capable of In-context Scheming

Alexander Meinke, Bronson Schoen, Jérémy Scheurer +3

Frontier models are increasingly trained and deployed as autonomous agent. One safety concern is that AI agents might covertly pursue misaligned goals, hiding their true capabiliti…

cs.CR2024

Towards evaluations-based safety cases for AI scheming

Mikita Balesni, Marius Hobbhahn, David Lindner +13

We sketch how developers of frontier AI systems could construct a structured rationale -- a 'safety case' -- that an AI system is unlikely to cause catastrophic outcomes through sc…

cs.AI2024

Analyzing Probabilistic Methods for Evaluating Agent Capabilities

Axel Højmark, Govind Pimpale, Arjun Panickssery +2

To mitigate risks from AI systems, we need to assess their capabilities accurately. This is especially difficult in cases where capabilities are only rarely displayed. Phuong et al…