AI Paper Atlas

Research map / Large language models / Alignment and preferences

A Regret Minimization Framework on Preference Learning in Large Language Models

Suhwan Kim, Taehyun Cho, Geon-Hyeong Kim, Yu Jin Kim, Youngsoo Jang, Moontae Lee, Jungwoo Lee

ICML 2026 · Large language models › Alignment and preferences · approach: Robust and calibrated DPO · 0 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
add masked-solution data augmentation (30K) (ablation)DPO on Qwen3-1.7B · GSM8Kaccuracy: 72.55 -> 77.71 (+5.16)Improved
add masked-solution data augmentation (30K) (ablation)DPO on Qwen3-4B · GSM8Kaccuracy: 85.22 -> 88.63 (+3.41)Improved
add masked-solution data augmentation (30K) (ablation)DPO on Qwen3-1.7B · MATHaccuracy: 52.14 -> 53.52 (+1.38)Improved
add masked-solution data augmentation (30K) (ablation)DPO on Qwen3-4B · MATHaccuracy: 63.70 -> 64.46 (+0.72)Improved
add masked-solution data augmentation (30K) (ablation)RePO on Qwen3-1.7B · GSM8Kaccuracy: 80.52 -> 80.29 (-0.23)No clear effect
add masked-solution data augmentation (30K) (ablation)RePO on Qwen3-4B · GSM8Kaccuracy: 90.60 -> 90.98 (+0.38)No clear effect
add masked-solution data augmentation (30K) (ablation)RePO on Qwen3-1.7B · MATHaccuracy: 54.50 -> 55.26 (+0.76)No clear effect
add masked-solution data augmentation (30K) (ablation)RePO on Qwen3-4B · MATHaccuracy: 65.54 -> 65.46 (-0.08)No clear effect
add masked augmentation to DPO lowers MSE to RePO scores (ablation)DPO on Qwen3-1.7B · token log-prob MSE vs RePOMSE: 0.073 -> 0.056Improved
replace DPO with RePOQwen3-1.7B-Base · AlpacaEval 2 LCLC win rate: 23.90 -> 36.61Improved
replace DPO with RePOQwen3-4B-Base · AlpacaEval 2 LCLC win rate: 32.89 -> 55.08Improved
replace DPO with RePOQwen3-4B-Base · GSM8Kaccuracy: 87.87 -> 90.60Improved

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