AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Improving Reasoning for Diffusion Language Models via Group Diffusion Policy Optimization

Kevin Rojas, Jiahe Lin, Kashif Rasul, Anderson Schneider, Yuriy Nevmyvaka, Molei Tao, Wei Deng

ICLR 2026 · Reinforcement learning › RL from feedback · approach: RL for LLM reasoning · 26 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 diffu-GRPO token-level likelihood with GDPO sequence-level ELBO (SFT first)LLaDA-8B-Instruct · Countdown (length 512)accuracy: 42.2 -> 80.86Improved
replace diffu-GRPO with GDPO (SFT first)LLaDA-8B-Instruct · GSM8K (length 512)accuracy: 82.1 -> 84.99Improved
replace diffu-GRPO with GDPO (SFT first)LLaDA-8B-Instruct · MATH500 (length 128)accuracy: 33.8 -> 33.2No clear effect
replace diffu-GRPO with GDPO (SFT first)LLaDA-8B-Instruct · Sudoku (length 512)accuracy: 9.5 -> 26.17Improved
replace diffu-GRPO with GDPOLLaDA-8B-Instruct · MBPP (length 256)pass rate: 45.5 -> 50.6Improved
replace diffu-GRPO with GDPOLLaDA-8B-Instruct · HumanEval (length 128)pass rate: 29.3 -> 26.2Hurt
replace diffu-GRPO with GDPOLLaDA-8B-Instruct · HumanEval (length 512)pass rate: 34.8 -> 39.0Improved
RL fine-tuning with GDPO on coding without SFTLLaDA-8B-Instruct · MBPP (length 256)pass rate: 41.2 -> 50.6Improved
use 3 quadrature points instead of 2 (ablation)GDPO · coding tasksaccuracyImproved
use semi-deterministic quadrature ELBO instead of naive Monte Carlo (ablation)GDPO · Countdowntest accuracyImproved
use GDPO instead of diffu-GRPO at longer generation lengthLLaDA-8B-Instruct · 512-token sequencesaccuracyImproved

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