Aligning Large Language Models by On-Policy Self-Judgment

• URL: http://arxiv.org/abs/2402.11253v3
• Date: Tue, 25 Jun 2024 13:39:52 GMT
• Title: Aligning Large Language Models by On-Policy Self-Judgment
• Authors: Sangkyu Lee, Sungdong Kim, Ashkan Yousefpour, Minjoon Seo, Kang Min Yoo, Youngjae Yu,
• Abstract summary: Existing approaches for aligning large language models with human preferences face a trade-off that requires a separate reward model (RM) for on-policy learning. We present a novel alignment framework, SELF-JUDGE, that does on-policy learning and is parameter efficient. We show that the rejecting sampling by itself can improve performance further without an additional evaluator.
• Score: 49.31895979525054
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Existing approaches for aligning large language models with human preferences face a trade-off that requires a separate reward model (RM) for on-policy learning. In this paper, we present a novel alignment framework, SELF-JUDGE that (1) does on-policy learning and 2) is parameter efficient, as it does not require an additional RM for evaluating the samples for on-policy learning. To this end, we propose Judge-augmented Supervised Fine-Tuning (JSFT) to train a single model to act as both a policy and a judge. Specifically, we view the pairwise judgment task, choosing the better response from a response pair, as a special case of the instruction-following task. The resulting model can judge preferences of on-the-fly responses from current policy initialized from itself. Experimental results show the efficacy of SELF-JUDGE, outperforming baselines in preference benchmarks. We also show that the rejecting sampling by itself can improve performance further without an additional evaluator.

Related papers

• Language Models are Few-Shot Graders [0.12289361708127876]
  We present an ASAG pipeline leveraging state-of-the-art LLMs. We compare the grading performance of three OpenAI models: GPT-4, GPT-4o, and o1-preview. Our findings indicate that providing graded examples enhances grading accuracy, with RAG-based selection outperforming random selection.
  arXiv  Detail & Related papers  (2025-02-18T23:38:21Z)
• HREF: Human Response-Guided Evaluation of Instruction Following in Language Models [61.273153125847166]
  We develop a new evaluation benchmark, Human Response-Guided Evaluation of Instruction Following (HREF) In addition to providing reliable evaluation, HREF emphasizes individual task performance and is free from contamination. We study the impact of key design choices in HREF, including the size of the evaluation set, the judge model, the baseline model, and the prompt template.
  arXiv  Detail & Related papers  (2024-12-20T03:26:47Z)
• A Systematic Examination of Preference Learning through the Lens of Instruction-Following [83.71180850955679]
  We use a novel synthetic data generation pipeline to generate 48,000 instruction unique-following prompts. With our synthetic prompts, we use two preference dataset curation methods - rejection sampling (RS) and Monte Carlo Tree Search (MCTS) Experiments reveal that shared prefixes in preference pairs, as generated by MCTS, provide marginal but consistent improvements. High-contrast preference pairs generally outperform low-contrast pairs; however, combining both often yields the best performance.
  arXiv  Detail & Related papers  (2024-12-18T15:38:39Z)
• Self-supervised Preference Optimization: Enhance Your Language Model with Preference Degree Awareness [27.43137305486112]
  We propose a novel Self-supervised Preference Optimization (SPO) framework, which constructs a self-supervised preference degree loss combined with the alignment loss. The results demonstrate that SPO can be seamlessly integrated with existing preference optimization methods to achieve state-of-the-art performance.
  arXiv  Detail & Related papers  (2024-09-26T12:37:26Z)
• Direct Judgement Preference Optimization [66.83088028268318]
  We train large language models (LLMs) as generative judges to evaluate and critique other models' outputs. We employ three approaches to collect the preference pairs for different use cases, each aimed at improving our generative judge from a different perspective. Our model robustly counters inherent biases such as position and length bias, flexibly adapts to any evaluation protocol specified by practitioners, and provides helpful language feedback for improving downstream generator models.
  arXiv  Detail & Related papers  (2024-09-23T02:08:20Z)
• TSO: Self-Training with Scaled Preference Optimization [14.3799656174528]
  We propose TSO, a framework for preference optimization that conducts self-training preference learning without training an additional reward model. TSO enhances the diversity of responses by constructing a model matrix and incorporating human preference responses. Experimental results demonstrate that TSO outperforms existing mainstream methods on various alignment evaluation benchmarks.
  arXiv  Detail & Related papers  (2024-08-31T05:37:01Z)
• Secrets of RLHF in Large Language Models Part II: Reward Modeling [134.97964938009588]
  We introduce a series of novel methods to mitigate the influence of incorrect and ambiguous preferences in the dataset. We also introduce contrastive learning to enhance the ability of reward models to distinguish between chosen and rejected responses.
  arXiv  Detail & Related papers  (2024-01-11T17:56:59Z)
• SALMON: Self-Alignment with Instructable Reward Models [80.83323636730341]
  This paper presents a novel approach, namely SALMON, to align base language models with minimal human supervision. We develop an AI assistant named Dromedary-2 with only 6 exemplars for in-context learning and 31 human-defined principles.
  arXiv  Detail & Related papers  (2023-10-09T17:56:53Z)
• Small Changes Make Big Differences: Improving Multi-turn Response Selection \\in Dialogue Systems via Fine-Grained Contrastive Learning [27.914380392295815]
  Retrieve-based dialogue response selection aims to find a proper response from a candidate set given a multi-turn context. We propose a novel textbfFine-textbfGrained textbfContrastive (FGC) learning method for the response selection task based on PLMs.
  arXiv  Detail & Related papers  (2021-11-19T11:07:07Z)

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.