AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Beyond Markovian: Reflective Exploration via Bayes-Adaptive RL for LLM Reasoning

Shenao Zhang, Yaqing Wang, Yinxiao Liu, Tianqi Liu, Peter Grabowski, Eugene Ie, Zhaoran Wang, Yunxuan Li

ICLR 2026 · Reinforcement learning › RL from feedback · approach: RL for LLM reasoning · 12 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 BARL instead of GRPOQwen2.5-Math-1.5B · six math benchmarks (average)accuracy (%): 51.0 -> 53.3Improved
use BARL instead of GRPOQwen2.5-Math-7B · six math benchmarks (average)accuracy (%): 57.1 -> 59.4Improved
use BARL instead of GRPODeepSeek-R1-Distill-Llama-8B · six math benchmarks (average)accuracy (%): 52.1 -> 52.7Improved
use BARL instead of progress-reward baselineQwen2.5-Math-1.5B · AIMEaccuracy (%): 14.4 -> 17.8Improved
use BARL instead of progress-reward baselineDeepSeek-R1-Distill-Llama-8B · MATH / GSM8Kaccuracy (%): GSM8K 85.9 -> 85.4No clear effect
use BARL over GRPO and progress for token efficiency (ablation)Qwen2.5-Math-1.5B · GSM8K and MATH pass@kaverage tokens: up to 39% fewer than progress, 50% fewer than GRPOImproved
add Bayes-adaptive hypothesis weighting (BARL) vs Markovian RL in didactic tasksmall transformer · synthetic repeat-token taskevaluation accuracyImproved
use more informative candidate set (|M|=3 repeated patterns) (ablation)BARL · synthetic repeat-token taskevaluation accuracyImproved
increase frequency of reflection behaviorsLLMs · math benchmarksaccuracyNo clear effect
restrict GRPO response length to 32 tokens (ablation)GRPO · training accuracytraining accuracyNo clear effect

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