6 papers
Infusion: Shaping Model Behavior by Editing Training Data via Influence Functions
J Rosser, Robert Kirk, Edward Grefenstette +2
Influence functions are commonly used to attribute model behavior to training documents. We explore the reverse: crafting training data that induces model behavior. Our framework,…
Improving Regret Approximation for Unsupervised Dynamic Environment Generation
Harry Mead, Bruno Lacerda, Jakob Foerster +1
Unsupervised Environment Design (UED) seeks to automatically generate training curricula for reinforcement learning (RL) agents, with the goal of improving generalisation and zero-…
DéjàQ: Open-Ended Evolution of Diverse, Learnable and Verifiable Problems
Willem Röpke, Samuel Coward, Andrei Lupu +3
Recent advances in reasoning models have yielded impressive results in mathematics and coding. However, most approaches rely on static datasets, which have been suggested to encour…
AI & Human Co-Improvement for Safer Co-Superintelligence
Jason Weston, Jakob Foerster
Self-improvement is a goal currently exciting the field of AI, but is fraught with danger, and may take time to fully achieve. We advocate that a more achievable and better goal fo…
Measuring what Matters: Construct Validity in Large Language Model Benchmarks
Andrew M. Bean, Ryan Othniel Kearns, Angelika Romanou +39
Evaluating large language models (LLMs) is crucial for both assessing their capabilities and identifying safety or robustness issues prior to deployment. Reliably measuring abstrac…
Learning to Reason at the Frontier of Learnability
Thomas Foster, Anya Sims, Johannes Forkel +2
Reinforcement learning is now widely adopted as the final stage of large language model training, especially for reasoning-style tasks such as maths problems. Typically, models att…