Updates
Our research
Evaluations, interpretability and governance work on deceptive and scheming behaviour in frontier AI systems.
Measuring Reward-Seeking via Contrastive Belief Updates
Visible forms of misbehavior are dropping in frontier models. Does that mean the models are becoming aligned, or are they just getting better at doing whatever they believe their grader rewards? Our paper finds that production reinforcement learning increases reward-seeking.
Stress Testing Deliberative Alignment for Anti-Scheming Training
We partnered with OpenAI to assess frontier language models for early signs of scheming — covertly pursuing misaligned goals — in controlled stress-tests (non-typical environments), and studied a training method that can significantly reduce (but not eliminate) these behaviors. Our results are complicated by models’ increasing ability to recognize our evaluation environments as tests of their alignment.
17 September 2025
Science of Scheming
Latest
All Publications
30 May 2024
The Local Interaction Basis: Identifying Computationally-Relevant and Sparsely Interacting Features in Neural Networks
Interpretability
Interpretability
30 May 2024
Identifying functionally important features with end-to-end sparse dictionary learning
Interpretability
Interpretability
09 November 2023
Large Language Models can Strategically Deceive their Users when Put Under Pressure
Evaluations
Evaluations