Towards a Theory of Learnability via Debate
cs.LG
Submitted: 2026-11-12
Updated: 2026-11-14
Terminology
Sources
- Concrete Problems in AI Safety
- Training Language Models to Win Debates with Self-Play Improves Judge Accuracy
- Avoiding Obfuscation with Prover-Estimator Debate
- The Hidden Game Problem
- Risks from Learned Optimization in Advanced Machine Learning Systems
- AI safety via debate
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis