paper-with-me

Papers

Latent Adversarial Regularization for Offline Preference Optimization

2026-01-29 · Enyi Jiang, Yibo Jacky Zhang, Yinglun Xu, Andreas Haupt, Nancy Amato, Sanmi Koyejo arxiv

Learning from human feedback typically relies on preference optimization that constrains policy updates through token-level regularization. However, preference optimization for language models is particularly challenging because token-space similarity does not imply semantic or behavioral similarity. To address this challenge, we leverage latent-space regularization for language model preference optimization. We introduce GANPO, which achieves latent-space regularization by penalizing divergence between the internal representations of a policy model and a reference model. Given that latent representations are not associated with explicit probability densities, we adopt an adversarial approach inspired by GANs to minimize latent-space divergence. We integrate GANPO as a regularizer into existing offline preference optimization objectives. Experiments across multiple model architectures and tasks show consistent improvements from latent-space regularization. Further, by comparing GANPO-induced inferential biases with those from token-level regularization, we find that GANPO provides more robust structural feedback under distributional shift and noise while maintaining comparable downstream performance with minor computational overhead.

📄 PDF Abstract BibTeX arXiv:2601.22083

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Generalized Preference Optimization: A Unified Approach to Offline Alignment

2024-02-08 · Yunhao Tang, Zhaohan Daniel Guo, Zeyu Zheng, Daniele Calandriello 외

Offline preference optimization allows fine-tuning large models directly from offline data, and has proved effective in recent alignment practices. We propose generalized preference optimization (GPO), a family of offlin…

Correcting the Mythos of KL-Regularization: Direct Alignment without Overoptimization via Chi-Squared Preference Optimization

2024-07-18 · Audrey Huang, Wenhao Zhan, Tengyang Xie, Jason D. Lee 외

Language model alignment methods, such as reinforcement learning from human feedback (RLHF), have led to impressive advances in language model capabilities, but existing techniques are limited by a widely observed phenom…

Language ModelingLanguage Modelling

The Importance of Online Data: Understanding Preference Fine-tuning via Coverage

2024-06-03 · Yuda Song, Gokul Swamy, Aarti Singh, J. Andrew Bagnell 외

Learning from human preference data has emerged as the dominant paradigm for fine-tuning large language models (LLMs). The two most common families of techniques -- online reinforcement learning (RL) such as Proximal Pol…

Reinforcement Learning (RL)

Failure Modes of Maximum Entropy RLHF

2025-09-24 · Ömer Veysel Çağatan, Barış Akgün arxiv

In this paper, we show that Simple Preference Optimization (SimPO) can be derived as Maximum Entropy Reinforcement Learning, providing a theoretical foundation for this reference-free method. Motivated by SimPO's strong …

Reinforcement Learning

Adversarial Policy Optimization for Offline Preference-based Reinforcement Learning

2025-03-07 · Hyungkyu Kang, Min-hwan Oh

In this paper, we study offline preference-based reinforcement learning (PbRL), where learning is based on pre-collected preference feedback over pairs of trajectories. While offline PbRL has demonstrated remarkable empi…

continuous-controlContinuous Controlreinforcement-learningReinforcement Learning