AI Paper Atlas

Research map / Reinforcement learning / Policy gradient and actor-critic

Revisiting Group Relative Policy Optimization: Insights into On-Policy and Off-Policy Training

Youssef Mroueh, Nicolas Dupuis, Brian Belgodere, Apoorva Nitsure, Mattia Rigotti, Kristjan Greenewald, Jiri Navratil, Jarret Ross, Jesus Rios

ICLR 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Model-based RL with guarantees · 39 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
mask zero-variance samples in on-policy GRPO (ablation)on-policy GRPO · GSM8K testPass@1 (%): 45 -> 50Improved
use off-policy GRPO (server update every v=10, i=1) instead of on-policyGRPO · GSM8K testPass@1 (%): 45 -> 50Improved
reuse fixed batch for 10 SGD iterations (v1,i10) off-policy variant (ablation)GRPO · GSM8K testconvergence speedHurt
use off-policy GRPO (v10,i1) instead of on-policyGRPO · Aime24Pass@1 mean: 0.3022 -> 0.3049No clear effect
use off-policy GRPO (v10,i1) instead of on-policyGRPO · Math500extractive match mean: 0.8519 -> 0.8474No clear effect
reset reference policy to latest model each stage (ablation)GRPO · GSM8K testPass@1Improved

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