Anchored Preference Optimization and Contrastive Revisions: Addressing Underspecification in Alignment
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.
Code (1)
Tasks
Contrastive LearningMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
RationalRewards: Reasoning Rewards Scale Visual Generation Both Training and Test Time
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 LearningToken Preference Optimization with Self-Calibrated Visual-Anchored Rewards for Hallucination Mitigation
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…
HallucinationBAPO: Base-Anchored Preference Optimization for Overcoming Forgetting in Large Language Models Personalization
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+1Self-Improving Robust Preference Optimization
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
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…