paper-with-me

Papers

Anchored Preference Optimization and Contrastive Revisions: Addressing Underspecification in Alignment

2024-08-12 · Karel D'Oosterlinck, Winnie Xu, Chris Develder, Thomas Demeester, Amanpreet Singh, Christopher Potts, Douwe Kiela, Shikib Mehri

Large Language Models (LLMs) are often aligned using contrastive alignment objectives and preference pair datasets. The interaction between model, paired data, and objective makes alignment a complicated procedure, sometimes producing subpar results. We study this and find that (i) preference data gives a better learning signal when the underlying responses are contrastive, and (ii) alignment objectives lead to better performance when they specify more control over the model during training. Based on these insights, we introduce Contrastive Learning from AI Revisions (CLAIR), a data-creation method which leads to more contrastive preference pairs, and Anchored Preference Optimization (APO), a controllable and more stable alignment objective. We align Llama-3-8B-Instruct using various comparable datasets and alignment objectives and measure MixEval-Hard scores, which correlate highly with human judgments. The CLAIR preferences lead to the strongest performance out of all datasets, and APO consistently outperforms less controllable objectives. Our best model, trained on 32K CLAIR preferences with APO, improves Llama-3-8B-Instruct by 7.65%, closing the gap with GPT4-turbo by 45%. Our code is available at https://github.com/ContextualAI/CLAIR_and_APO.

📄 PDF Abstract BibTeX arXiv:2408.06266

Code (1)

contextualai/clair_and_apo 공식 구현

Tasks

Contrastive Learning

Methods 이 논문이 사용한 방법론

Contrastive Learning 설명 없음
ALIGN In the ALIGN method, visual and language representations are jointly trained from noisy image alt-text data. The image and text encoders are learned via contrastive loss…

Similar Papers 제목 키워드 기반

RationalRewards: Reasoning Rewards Scale Visual Generation Both Training and Test Time

2026-04-13 · Haozhe Wang, Cong Wei, Weiming Ren, Jiaming Liu 외 arxiv

Most reward models for visual generation reduce rich human judgments to a single unexplained score, discarding the reasoning that underlies preference. We show that teaching reward models to produce explicit, multi-dimen…

Reinforcement Learning

Token Preference Optimization with Self-Calibrated Visual-Anchored Rewards for Hallucination Mitigation

2024-12-19 · Jihao Gu, Yingyao Wang, Meng Cao, Pi Bu 외

Direct Preference Optimization (DPO) has been demonstrated to be highly effective in mitigating hallucinations in Large Vision Language Models (LVLMs) by aligning their outputs more closely with human preferences. Despit…

Hallucination

BAPO: Base-Anchored Preference Optimization for Overcoming Forgetting in Large Language Models Personalization

2024-06-30 · Gihun Lee, Minchan Jeong, Yujin Kim, Hojung Jung 외

While learning to align Large Language Models (LLMs) with human preferences has shown remarkable success, aligning these models to meet the diverse user preferences presents further challenges in preserving previous know…

Continual LearningGeneral KnowledgeKnowledge DistillationPersonalized and Emotional Conversation+1

Self-Improving Robust Preference Optimization

2024-06-03 · Eugene Choi, Arash Ahmadian, Matthieu Geist, Oilvier Pietquin 외

Online and offline RLHF methods, such as PPO and DPO, have been highly successful in aligning AI with human preferences. Despite their success, however, these methods suffer from fundamental limitations: (a) Models train…

No Hidden Prompts Needed! You Can Game AI Peer Review with Presentation-Only Revisions

2026-06-11 · Xu Yang, Zhizhou Sha, Junbo Li, Jian Yu 외 arxiv

As AI-generated reviews move from experimental tools into peer-review infrastructure, most robustness concerns have focused on explicit attacks such as hidden instructions and prompt injection. We study a harder and more…