paper-with-me

Papers

RePO: ReLU-based Preference Optimization

2025-03-10 · Junkang Wu, Kexin Huang, Xue Wang, Jinyang Gao, Bolin Ding, Jiancan Wu, Xiangnan He, Xiang Wang

Aligning large language models (LLMs) with human preferences is critical for real-world deployment, yet existing methods like RLHF face computational and stability challenges. While DPO establishes an offline paradigm with single hyperparameter $\beta$, subsequent methods like SimPO reintroduce complexity through dual parameters ($\beta$, $\gamma$). We propose {ReLU-based Preference Optimization (RePO)}, a streamlined algorithm that eliminates $\beta$ via two advances: (1) retaining SimPO's reference-free margins but removing $\beta$ through gradient analysis, and (2) adopting a ReLU-based max-margin loss that naturally filters trivial pairs. Theoretically, RePO is characterized as SimPO's limiting case ($\beta \to \infty$), where the logistic weighting collapses to binary thresholding, forming a convex envelope of the 0-1 loss. Empirical results on AlpacaEval 2 and Arena-Hard show that RePO outperforms DPO and SimPO across multiple base models, requiring only one hyperparameter to tune.

📄 PDF Abstract BibTeX arXiv:2503.07426

Code (1)

junkangwu/repo 공식 구현 pytorch

Methods 이 논문이 사용한 방법론

DPO 설명 없음
BASE 설명 없음

Similar Papers 제목 키워드 기반

Neural Characteristic Activation Analysis and Geometric Parameterization for ReLU Networks

2023-05-25 · Wenlin Chen, Hong Ge

We introduce a novel approach for analyzing the training dynamics of ReLU networks by examining the characteristic activation boundaries of individual ReLU neurons. Our proposed analysis reveals a critical instability in…

Stochastic Optimization

Radiology Report Generation via Multi-objective Preference Optimization

2024-12-12 · Ting Xiao, Lei Shi, Peng Liu, Zhe Wang 외

Automatic Radiology Report Generation (RRG) is an important topic for alleviating the substantial workload of radiologists. Existing RRG approaches rely on supervised regression based on different architectures or additi…

Multi-Objective Reinforcement LearningReinforcement Learning (RL)

$i$REPO: $i$mplicit Reward Pairwise Difference based Empirical Preference Optimization

2024-05-24 · Long Tan Le, Han Shu, Tung-Anh Nguyen, Choong Seon Hong 외

While astonishingly capable, large Language Models (LLM) can sometimes produce outputs that deviate from human expectations. Such deviations necessitate an alignment phase to prevent disseminating untruthful, toxic, or b…

Language Model EvaluationLanguage ModelingLanguage Modellingregression

EraseReLU: A Simple Way to Ease the Training of Deep Convolution Neural Networks

2017-09-22 · Xuanyi Dong, Guoliang Kang, Kun Zhan, Yi Yang

For most state-of-the-art architectures, Rectified Linear Unit (ReLU) becomes a standard component accompanied with each layer. Although ReLU can ease the network training to an extent, the character of blocking negative…

BlockingImage Classification

ReLU's Revival: On the Entropic Overload in Normalization-Free Large Language Models

2024-10-12 · Nandan Kumar Jha, Brandon Reagen

LayerNorm is a critical component in modern large language models (LLMs) for stabilizing training and ensuring smooth optimization. However, it introduces significant challenges in mechanistic interpretability, outlier f…

Decoder