English

What is the objective of reasoning with reinforcement learning?

Machine Learning 2025-10-16 v1 Optimization and Control

Abstract

We show that several popular algorithms for reinforcement learning in large language models with binary rewards can be viewed as stochastic gradient ascent on a monotone transform of the probability of a correct answer given a prompt. In particular, the transformation associated with rejection sampling algorithms is the logarithm and that associated with the GRPO algorithm is the arcsine of the square root.

Keywords

Cite

@article{arxiv.2510.13651,
  title  = {What is the objective of reasoning with reinforcement learning?},
  author = {Damek Davis and Benjamin Recht},
  journal= {arXiv preprint arXiv:2510.13651},
  year   = {2025}
}