Iterative Length-Regularized Direct Preference Optimization: A Case Study on Improving 7B Language Models to GPT-4 Level

• URL: http://arxiv.org/abs/2406.11817v1
• Date: Mon, 17 Jun 2024 17:55:38 GMT
• Title: Iterative Length-Regularized Direct Preference Optimization: A Case Study on Improving 7B Language Models to GPT-4 Level
• Authors: Jie Liu, Zhanhui Zhou, Jiaheng Liu, Xingyuan Bu, Chao Yang, Han-Sen Zhong, Wanli Ouyang,
• Abstract summary: We show that iLR-DPO can enhance a 7B model to perform on par with GPT-4 without increasing verbosity. Specifically, our 7B model achieves a $50.5%$ length-controlled win rate against $texttGPT-4 Preview$ on AlpacaEval 2.0.
• Score: 50.897438358317686
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Direct Preference Optimization (DPO), a standard method for aligning language models with human preferences, is traditionally applied to offline preferences. Recent studies show that DPO benefits from iterative training with online preferences labeled by a trained reward model. In this work, we identify a pitfall of vanilla iterative DPO - improved response quality can lead to increased verbosity. To address this, we introduce iterative length-regularized DPO (iLR-DPO) to penalize response length. Our empirical results show that iLR-DPO can enhance a 7B model to perform on par with GPT-4 without increasing verbosity. Specifically, our 7B model achieves a $50.5\%$ length-controlled win rate against $\texttt{GPT-4 Preview}$ on AlpacaEval 2.0, and excels across standard benchmarks including MT-Bench, Arena-Hard and OpenLLM Leaderboard. These results demonstrate the effectiveness of iterative DPO in aligning language models with human feedback.

Related papers

• GDPO: Learning to Directly Align Language Models with Diversity Using GFlowNets [19.485572131953937]
  We propose a practical application of a diversity-seeking RL algorithm called GFlowNet-DPO (GDPO) in an offline preference alignment setting. Empirical results show GDPO can generate far more diverse responses than the baseline methods.
  arXiv  Detail & Related papers  (2024-10-19T13:07:52Z)
• TPO: Aligning Large Language Models with Multi-branch & Multi-step Preference Trees [14.84379332031731]
  We introduce Tree Preference Optimization (TPO), that does not sample paired preference responses from the preference tree. TPO formulates the language model alignment as a Preference List Ranking problem, where the policy can learn more effectively from a ranked preference list of responses.
  arXiv  Detail & Related papers  (2024-10-10T22:22:05Z)
• General Preference Modeling with Preference Representations for Aligning Language Models [51.14207112118503]
  We introduce preference representation learning, an approach that embeds responses into a latent space to capture intricate preference structures efficiently. We also propose preference score-based General Preference Optimization (GPO), which generalizes reward-based reinforcement learning from human feedback. Our method may enhance the alignment of foundation models with nuanced human values.
  arXiv  Detail & Related papers  (2024-10-03T04:22:55Z)
• Bootstrapping Language Models with DPO Implicit Rewards [45.68366127605774]
  Direct preference optimization (DPO) has greatly simplified the process from past work in reinforcement learning from human feedback. In this work, we make a novel observation that this implicit reward model can by itself be used in a bootstrapping fashion to further align the LLM. Our approach, named self-alignment with DPO ImpliCit rEwards (DICE), shows great improvements in alignment and achieves superior performance.
  arXiv  Detail & Related papers  (2024-06-14T06:57:18Z)
• Self-Augmented Preference Optimization: Off-Policy Paradigms for Language Model Alignment [104.18002641195442]
  We introduce Self-Augmented Preference Optimization (SAPO), an effective and scalable training paradigm that does not require existing paired data. Building on the self-play concept, which autonomously generates negative responses, we further incorporate an off-policy learning pipeline to enhance data exploration and exploitation.
  arXiv  Detail & Related papers  (2024-05-31T14:21:04Z)
• Self-Play Preference Optimization for Language Model Alignment [75.83359213697854]
  Recent advancements suggest that directly working with preference probabilities can yield a more accurate reflection of human preferences. We propose a self-play-based method for language model alignment, which treats the problem as a constant-sum two-player game. Our approach, dubbed Self-Play Preference Optimization (SPPO), utilizes iterative policy updates to provably approximate the Nash equilibrium.
  arXiv  Detail & Related papers  (2024-05-01T17:59:20Z)
• Filtered Direct Preference Optimization [7.060398061192042]
  Reinforcement learning from human feedback (RLHF) plays a crucial role in aligning language models with human preferences. This paper addresses the issue of text quality within the preference dataset by focusing on direct preference optimization (DPO) We propose an extension of DPO, termed filtered direct preference optimization (fDPO)
  arXiv  Detail & Related papers  (2024-04-22T03:05:19Z)
• Direct Preference Optimization with an Offset [58.7977683502207]
  Direct preference optimization (DPO) is a successful strategy for aligning large language models with human preferences. We propose a generalization of DPO, termed DPO with an offset (ODPO), that does not treat every preference pair equally during fine-tuning.
  arXiv  Detail & Related papers  (2024-02-16T10:55:38Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.