AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Stabilizing Policy Gradients for Sample-Efficient Reinforcement Learning in LLM Reasoning

Luckeciano Carvalho Melo, Alessandro Abate, Yarin Gal

ICLR 2026 · Reinforcement learning › RL from feedback · approach: GRPO stability and variants · 5 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use aggressive update regime (5x learning rate, 12x smaller batch) with standard GRPO (ablation)GRPO · MATH / TESTaccuracyMixed
add curvature-aware token selection (CAPO) to GRPO under aggressive updatesGRPO (A) · MATHcompletions needed / stability: 30x fewer completions on MATHImproved
add curvature-aware token selection (CAPO) to GRPO under aggressive updatesGRPO (A) · TEST (8 benchmarks avg)completions needed / stability: 9x fewer completionsImproved
add curvature-aware selection to Dr.GRPO (Dr.CAPO) (ablation)Dr.GRPO · MATHaccuracy / collapseImproved
add curvature-aware selection to REINFORCE (ReinCAPO) (ablation)REINFORCE · MATHaccuracy / collapseImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy