AI Paper Atlas

Research map / Large language models / Alignment and preferences

RMO: Towards Better LLM Alignment via Reshaping Reward Margin Distributions

Yanchi Ru, Yue Huang, Xiangliang Zhang

AAAI 2026 · Large language models › Alignment and preferences · approach: DPO design choices · 1 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
build batches that maximize intra-batch reward margin variance (High-Var) instead of random (ablation)DPO · AlpacaEval 2.0 on Anthropic HH, Llama-3.2-1B-InstructLC win rate (%): 54.00 -> 56.39Improved
build batches that minimize margin variance (Low-Var) instead of random (ablation)DPO · AlpacaEval 2.0 on Anthropic HH, Llama-3.2-1B-InstructLC win rate (%): 54.00 -> 49.67Hurt
add Dual Denoising Filtering (6% filter ratio) (ablation)DPO with random batches · AlpacaEval 2.0 on Anthropic HH, Llama-3.2-1B-InstructLC win rate (%): 54.00 -> 56.05Improved
add Pairwise Margin Amplification (lambda_reg=2) (ablation)DPO with high-variance batches · AlpacaEval 2.0 on Anthropic HH, Pythia-1.4BLC win rate (%): 59.20 -> 62.34Improved
add Dual Denoising Filtering to high-variance DPO (ablation)DPO with High-Var batches · AlpacaEval 2.0 on Anthropic HH, Llama-3.2-1B-InstructLC win rate (%): 56.39 -> 56.43No clear effect
add full RMO to high-variance DPODPO · AlpacaEval 2.0 on Anthropic HH, Llama-3.2-1B-InstructLC win rate (%): 56.39 -> 59.78Improved
add RMO to SimPOSimPO · AlpacaEval 2.0 on UltraFeedback, Llama-3.2-3B-InstructLC win rate (%): 65.91 -> 69.52Improved
add Pairwise Margin Amplification to RMO on SimPO (ablation)RMO (SimPO) · AlpacaEval 2.0 on UltraFeedback, Llama-3.2-3B-InstructLC win rate (%): 68.32 -> 69.52Improved
add Batch Margin Diversification to SimPO (ablation)SimPO · AlpacaEval 2.0 on Anthropic HH, Llama-3.2-3B-InstructLC win rate (%): 60.00 -> 61.44Mixed
increase lambda_reg beyond 2 (ablation)Pairwise Margin Amplification · AlpacaEval 2.0, Llama-3.2-1B-InstructLC and WRHurt
increase lambda_reg (ablation)Pairwise Margin Amplification · Preference accuracy, Llama-3.2-1B-InstructaccuracyImproved
increase filter ratio beyond 10% (ablation)Dual Denoising Filtering · AlpacaEval 2.0, Llama-3.2-1B-InstructLC win rate (%)Hurt

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