AI Paper Atlas

Research map / Large language models / Reasoning and chain of thought

Learning to Reason in Structured In-context Environments with Reinforcement Learning

Peng Yu, Zeyuan Zhao, Shao Zhang, Luoyi Fu, Xinbing Wang, Ying Wen

ICLR 2026 · Large language models › Reasoning and chain of thought · approach: RL with verifiable rewards for reasoning

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes learning to reason in structured in-context environments with RL, where the abstract argues an ideal LLM reasoning environment should be scalable and general (full text unavailable, so details are limited).

Problem

RL finetuning of LLMs for reasoning depends on the environment, and an ideal reasoning environment should be scalable and general; the abstract is truncated so the specific gap is not fully stated.

Approach

Uses structured in-context environments as the RL training environment for LLM reasoning; the specific design is not stated in the available abstract.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add structured in-context environment during RL (ablation)GRPO on Qwen2.5-7B-Instruct · WebQSPaccuracy: 59.7 -> 93.4Improved
add structured in-context environment during RL (ablation)GRPO on Qwen2.5-7B-Instruct · GrailQAaccuracy: 20.8 -> 85.8Improved
use RL in SIE instead of SFT on distilled structured reasoning dataQwen2.5-7B-Instruct · WebQSPaccuracy: 40.5 -> 93.4Improved
RL fine-tuning in SIE vs CoT prompting for out-of-domain mathQwen2.5-7B-Instruct · GSM8Kaccuracy: 29.2 -> 87.4Improved
RL fine-tuning in SIE vs CoT on out-of-domain logicQwen2.5-7B-Instruct · KK-hardaccuracy: 19.5 -> 29.0Improved
remove all supporting triples (SIE-0% vs SIE-100%) (ablation)RL w/ SIE Qwen2.5-7B-Instruct · WebQSPaccuracy: 93.4 -> 72.8Hurt
remove supporting triples (SIE-0% vs SIE-100%) (ablation)RL w/ SIE Qwen2.5-7B-Instruct · GSM8K (OOD)accuracy: 87.4 -> 87.1No clear effect
use PPO instead of GRPO (ablation)RL w/ SIE Qwen2.5-7B-Instruct · WebQSPaccuracy: 93.4 -> 85.4Hurt
use REINFORCE++ instead of GRPO (ablation)RL w/ SIE Qwen2.5-7B-Instruct · WebQSPaccuracy: 93.4 -> 93.1No clear effect
start RL from SFT checkpoint (ablation)RL w/ SIE Qwen2.5-7B-Instruct · WebQSPaccuracy: 93.4 -> 88.5 (WebQSP); 29.0 -> 33.5 (KK-hard)Mixed
add correctness reward instead of format-only reward (ablation)RL w/ SIE Qwen2.5-7B-Instruct · WebQSPaccuracy: 31.6 -> 93.4Improved
replace correctness reward with random noise (ablation)RL w/ SIE Qwen2.5-7B-Instruct · WebQSPaccuracy: 26.3 -> 6.3Hurt

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