AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Critique-RL: Training Critiquing Language Models Through Two-Stage RL for Improved Discrimination and Constructive Feedback

Zhiheng Xi, Jixuan Huang, Xin Guo, Boyang Hong, Dingwen Yang, Xiaoran Fan, Shuo Li, Zehui Chen, Junjie Ye, Siyu Yuan, Zhengyin Du, Xuesong Yao and 6 more

ICLR 2026 · Reinforcement learning › RL from feedback · approach: RL for LLM reasoning

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
add Stage I discrimination training (ablation)Critique-RL · MATH (Qwen2.5-3B)Acc@Refine: 45.9 -> 48.6Improved
add Stage II helpfulness training (ablation)Critique-RL · MATH (Qwen2.5-3B)Acc@Dis: 78.7 -> 82.8Improved
add Stage I (ablation)Critique-RL · AQuA (Qwen2.5-3B)Acc@Refine: 53.9 -> 56.7Improved
keep discrimination reward and KL regularization in Stage II (ablation)Critique-RL · MATH (Qwen2.5-3B)Acc@Dis: 77.7 -> 82.8Improved
keep discrimination reward and KL in Stage II (ablation)Critique-RL · AQuA (Qwen2.5-3B)Acc@Dis: 61.6 -> 69.9Improved
use rrefine reward instead of r_delta in Stage II (ablation)Critique-RL · MATH (Qwen2.5-3B)Acc@Refine: 48.2 -> 48.6Improved
use rrefine reward instead of rcorrection in Stage II (ablation)Critique-RL · MATH (Qwen2.5-3B)Acc@Refine: 47.7 -> 48.6Improved
add a second training iteration (ablation)Critique-RL · MATH (Qwen2.5-3B)Acc: 48.6 -> 51.0Improved
use online RL critic instead of SFT criticCritique-RL vs SFT · AQuA (Qwen2.5-3B)Acc: 46.46 -> 56.69Improved
use Critique-RL critic instead of Directly RL actorQwen2.5-7B · MATHAcc: 49.78 -> 58.40Improved
use Critique-RL instead of SCoReQwen2.5-7B · MATHAcc@Dis: 72.51 -> 85.20Improved
change beta from 0.01 to 0.012 (ablation)Critique-RL · GSM8K (Qwen2.5-3B)Acc: 75.89 -> 74.22No 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