AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Single-stream Policy Optimization

Zhongwen Xu, Zihan Ding

ICLR 2026 · Reinforcement learning › RL from feedback · approach: GRPO stability and variants · 33 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
replace GRPO with SPOQwen3-8B tool-integrated RL · Average of five benchmarksmaj@32: 60.4 -> 63.8Improved
replace GRPO with SPOQwen3-8B tool-integrated RL · BRUMO 25maj@32: 56.7 -> 64.0Improved
replace GRPO with SPOQwen3-8B tool-integrated RL · AIME 25maj@32: 72.1 -> 76.5Improved
replace GRPO with SPOQwen3-8B tool-integrated RL · HMMT 25maj@32: 44.2 -> 47.5Improved
replace GRPO with SPOQwen3-8B tool-integrated RL · AIME 24avg@32: 77.6 -> 74.9Hurt
replace GRPO with SPOQwen3-8B tool-integrated RL · Average of five benchmarksavg@32: 55.7 -> 56.0No clear effect
use group-free collection of the first 24 finished samples from 48 instead of group-based collectionagentic rollout simulation · Agentic batch-collection simulationtraining throughput speedup: 486s -> 112s (4.35x)Improved
add the adaptive KL-based value tracker baseline to SPO (vs no baseline) (ablation)SPO · AIME 25 (avg@16, ablation setting)avg@16Improved
use adaptive tracker baseline instead of static A*-PO baseline (ablation)SPO · AIME 25 (avg@16, ablation setting)avg@16Improved
add offline initialization of the value tracker (ablation)SPO · AIME 25 (avg@16, ablation setting)avg@16Improved
use SPO history-informed baseline instead of raw rewards (ablation)SPO · Advantage variance analysisadvantage variance: nearly 50% variance reductionImproved
use SPO instead of GRPO regarding ineffective samplesGRPO · Training dynamicsshare of zero-advantage samples: roughly 60% to over 80% (GRPO degenerate groups)Improved

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