Score: 1

Adaptive-Boundary-Clipping GRPO: Ensuring Bounded Ratios for Stable and Generalizable Training

Published: January 7, 2026 | arXiv ID: 2601.03895v1

By: Chi Liu, Xin Chen

Potential Business Impact:

Makes AI better at math and learning.

Business Areas:
A/B Testing Data and Analytics

Group Relative Policy Optimization (GRPO) has emerged as a popular algorithm for reinforcement learning with large language models (LLMs). However, upon analyzing its clipping mechanism, we argue that it is suboptimal in certain scenarios. With appropriate modifications, GRPO can be significantly enhanced to improve both flexibility and generalization. To this end, we propose Adaptive-Boundary-Clipping GRPO (ABC-GRPO), an asymmetric and adaptive refinement of the original GRPO framework. We demonstrate that ABC-GRPO achieves superior performance over standard GRPO on mathematical reasoning tasks using the Qwen3 LLMs. Moreover, ABC-GRPO maintains substantially higher entropy throughout training, thereby preserving the model's exploration capacity and mitigating premature convergence. The implementation code is available online to ease reproducibility https://github.com/chi2liu/ABC-GRPO.

Repos / Data Links

Page Count
10 pages

Category
Computer Science:
Machine Learning (CS)