AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Conditional Equivalence of DPO and RLHF: Assumptions, Failure Modes, and Provable Alignment

Zhiqin Yang, Yonggang Zhang, Wei Xue, Dong Fang, Bo Han, Yike Guo

ICML 2026 · Reinforcement learning › RL from feedback · approach: Reward modeling versus direct optimization

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 adaptive reference-based margin (CPO) to DPODPO, Llama-3-8B-Instruct · AlpacaEval 2LC win rate (%): 25.09 -> 26.57Improved
add adaptive reference-based margin (CPO) to DPODPO, Llama-3-8B-Instruct · Arena-Hardwin rate (%): 28.9 -> 32.6Improved
use CPO instead of DPO when the reference policy is misaligned (R=0.2)DPO, corrupted SFT reference · AlpacaEval 2 (R=0.2)LC win rate (%): 17.23 -> 27.60Improved
use CPO instead of DPO when the reference policy is misaligned (R=0.4)DPO, corrupted SFT reference · AlpacaEval 2 (R=0.4)LC win rate (%): 15.98 -> 24.34Improved
increase reference corruption ratio from 0.2 to 0.4 (ablation)CPO · AlpacaEval 2LC win rate (%): 27.60 -> 24.34Hurt
raise margin weight gamma from 0.15 to 0.20 (ablation)CPO · AlpacaEval 2 (gpt-4.1 annotator)LC win rate (%): 26.32 -> 32.56Improved
vary gamma within 0.2 to 0.4 (ablation)CPO · AlpacaEval 2 (gpt-4.1 annotator)WR and LC: WR 26-28%, LC 31-34%No clear effect
replace the adaptive margin with clipping the reference log-ratio to non-negative (ablation)CPO · AlpacaEval 2 (gpt-4.1 annotator)LC win rate (%): 33.97 -> 23.86Hurt
use CPO instead of DPODPO, Llama-3-8B-Instruct · IFEvalstrict accuracy (%): 34.01% -> 35.12%Improved
use CPO instead of DPODPO, Llama-3-8B-Instruct · IFEvalloose accuracy (%): 40.67% -> 43.99%Improved

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