Papers › Training Language Models to Win Debates with Self-Play Improves Judge Accuracy

Training Language Models to Win Debates with Self-Play Improves Judge Accuracy

25 Sep 2024arXiv:2409.16636archive 2025-07-28

Samuel Arnesen, David Rein, Julian Michael

We test the robustness of debate as a method of scalable oversight by training models to debate with data generated via self-play. In a long-context reading comprehension task, we find that language model based evaluators answer questions more accurately when judging models optimized to win debates. By contrast, we find no such relationship for consultancy models trained to persuade a judge without an opposing debater present. In quantitative and qualitative comparisons between our debate models and novel consultancy baselines, we find evidence that debate training encourages stronger and more informative arguments, showing promise that it can help provide high-quality supervision for tasks that are difficult to directly evaluate.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

samuelarnesen/nyu-debate-modeling officialmentioned in paperpytorchMIT report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Language ModelingLanguage ModellingReading Comprehension

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections