paper-with-me

홈 › Papers

Refined Direct Preference Optimization with Synthetic Data for Behavioral Alignment of LLMs

2024-02-12 · Víctor Gallego

In this paper, we introduce \emph{refined Direct Preference Optimization} (rDPO), a method for improving the behavioral alignment of Large Language Models (LLMs) without the need for human-annotated data. The method involves creating synthetic data using self-critique prompting by a teacher LLM and then utilising a generalized DPO loss function to distil to a student LLM. The loss function incorporates an additional external reward model to improve the quality of synthetic data, making rDPO robust to potential noise in the synthetic dataset. rDPO is shown to be effective in a diverse set of behavioural alignment tasks, such as improved safety, robustness against role-playing, and reduced sycophancy. Code to be released at https://github.com/vicgalle/refined-dpo.

📄 PDF Abstract BibTeX arXiv:2402.08005

Code (1)

vicgalle/refined-dpo 공식 구현

Methods 이 논문이 사용한 방법론

DPO 설명 없음
SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Similar Papers 제목 키워드 기반

Fine-Tuning Diffusion Generative Models via Rich Preference Optimization

2025-03-13 · Hanyang Zhao, Haoxian Chen, Yucheng Guo, Genta Indra Winata 외

We introduce Rich Preference Optimization (RPO), a novel pipeline that leverages rich feedback signals to improve the curation of preference pairs for fine-tuning text-to-image diffusion models. Traditional methods, like…

Self-alignment of Large Video Language Models with Refined Regularized Preference Optimization

2025-04-16 · Pritam Sarkar, Ali Etemad

Despite recent advances in Large Video Language Models (LVLMs), they still struggle with fine-grained temporal understanding, hallucinate, and often make simple mistakes on even simple video question-answering tasks, all…

HallucinationQuestion AnsweringVideo Question AnsweringVideo Understanding

Direct Value Optimization: Improving Chain-of-Thought Reasoning in LLMs with Refined Values

2025-02-19 · Hongbo Zhang, Han Cui, Guangsheng Bao, Linyi Yang 외

We introduce Direct Value Optimization (DVO), an innovative reinforcement learning framework for enhancing large language models in complex reasoning tasks. Unlike traditional methods relying on preference labels, DVO ut…

Knowledge Editing in Language Models via Adapted Direct Preference Optimization

2024-06-14 · Amit Rozner, Barak Battash, Lior Wolf, Ofir Lindenbaum

Large Language Models (LLMs) can become outdated over time as they may lack updated world knowledge, leading to factual knowledge errors and gaps. Knowledge Editing (KE) aims to overcome this challenge using weight updat…

knowledge editingWorld Knowledge

VERTIGO: Visual Preference Optimization for Cinematic Camera Trajectory Generation

2026-04-02 · Mengtian Li, Yuwei Lu, Feifei Li, Chenqi Gan 외 arxiv

Cinematic camera control relies on a tight feedback loop between director and cinematographer, where camera motion and framing are continuously reviewed and refined. Recent generative camera systems can produce diverse, …

Semantic Similarity