Selective-Advantage Entropy-Adaptive Horizon GRPO: Asymmetric Token-Level Discounting for Efficient Reinforcement Learning of Language Models
Group Relative Policy Optimisation (GRPO) has emerged as an effective reinforcement-learning algorithm for aligning language models on reasoning tasks, but it treats every token position and every sampled rollout symmetrically. We introduce two complementary extensions: (i) Adaptive-Horizon GRPO (AH-GRPO), which weights each token's policy gradient using a cumulative entropy-based discount that reduces the effective horizon when the model is uncertain, and (ii) Selective-Advantage AH-GRPO (SA-AH-GRPO), which applies this discounting only to negative-advantage rollouts, leaving positive-advantage, successful trajectories unattenuated. We evaluate standard GRPO with alpha = 0, AH-GRPO with ...
| Authors | Chirag Chawla, Rohan Charudatt Salvi, Madhav S. Baidya |
|---|---|
| Primary category | cs.LG |
| Published | 3 June 2026 |
| Last updated at source | 3 June 2026 |
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.