AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

ARLArena: A Unified Framework for Stable Agentic Reinforcement Learning

Xiaoxuan Wang, Han Zhang, Haixin Wang, Yidan Shi, Ruoyan Li, Kaiqiao Han, Chenyi Tong, Haoran Deng, Alexander K Taylor, Renliang Sun, Yanqiao Zhu, Jason Cong and 2 more

ICML 2026 · Reinforcement learning › RL from feedback · approach: Scaling RL with verifiable rewards · 0 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 sequence-level clipping (GSPO) instead of token-level GRPO clipping (ablation)GRPO · Avg over tasksscore: 46.16 -> 52.28Improved
use tolerant clipping (CISPO/SAPO) instead of hard clipping (ablation)GRPO · ALFWorldsuccess rate (%): 62.36 -> 54.42 (CISPO), 25.16 (SAPO)Hurt
add sequence masking of negative-advantage low-ratio sequences (ablation)CISPO · ALFWorldsuccess rate (%): 54.42 -> 78.88Improved
add sequence masking of negative-advantage low-ratio sequences (ablation)SAPO · ALFWorldsuccess rate (%): 25.16 -> 76.92Improved
increase KL coefficient to 0.05 (ablation)CISPO · ALFWorldsuccess rate (%): 54.42 -> 38.46Hurt
increase KL coefficient to 0.05 (ablation)SAPO · ALFWorldsuccess rate (%): 25.16 -> 48.05Improved
enlarge off-policy mini-update batch to 1024 (ablation)CISPO · ALFWorldsuccess rate (%): 54.42 -> 21.59Hurt
add fine-grained environment-level advantage (GiGPO) (ablation)GRPO · ALFWorldsuccess rate (%): 62.36 -> 81.09Improved
add entropy-based advantage reshaping (EMPG) (ablation)GRPO · ALFWorld / WebShopsuccess rate (%): 57.91 vs 62.36 on ALFWorld; 64.32 vs 57.71 on WebShopMixed
add dynamic filtering (ablation)GRPO · ALFWorldsuccess rate (%): 62.36 -> 49.58Hurt
add dynamic filtering (ablation)GIGPO · WebShopsuccess rate (%): 56.55 -> 76.82Improved
use seq-mean-token-mean loss aggregation (ablation)GRPO · Avg over tasksscore: 46.16 -> 39.23Hurt

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