Better Instruction-Following Through Minimum Bayes Risk

• URL: http://arxiv.org/abs/2410.02902v3
• Date: Mon, 28 Oct 2024 17:22:43 GMT
• Title: Better Instruction-Following Through Minimum Bayes Risk
• Authors: Ian Wu, Patrick Fernandes, Amanda Bertsch, Seungone Kim, Sina Pakazad, Graham Neubig,
• Abstract summary: General-purpose LLM judges capable of human-level evaluation provide a scalable and accurate way of evaluating instruction-following LLMs. One promising way of leveraging LLM judges for supervision is through Minimum Bayes Risk (MBR) decoding. MBR decoding uses a reference-based evaluator to select a high-quality output from amongst a set of candidate outputs.
• Score: 48.879360919760074
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: General-purpose LLM judges capable of human-level evaluation provide not only a scalable and accurate way of evaluating instruction-following LLMs but also new avenues for supervising and improving their performance. One promising way of leveraging LLM judges for supervision is through Minimum Bayes Risk (MBR) decoding, which uses a reference-based evaluator to select a high-quality output from amongst a set of candidate outputs. In the first part of this work, we explore using MBR decoding as a method for improving the test-time performance of instruction-following LLMs. We find that MBR decoding with reference-based LLM judges substantially improves over greedy decoding, best-of-N decoding with reference-free judges and MBR decoding with lexical and embedding-based metrics on AlpacaEval and MT-Bench. These gains are consistent across LLMs with up to 70B parameters, demonstrating that smaller LLM judges can be used to supervise much larger LLMs. Then, seeking to retain the improvements from MBR decoding while mitigating additional test-time costs, we explore iterative self-training on MBR-decoded outputs. We find that self-training using Direct Preference Optimisation leads to significant performance gains, such that the self-trained models with greedy decoding generally match and sometimes exceed the performance of their base models with MBR decoding.

Related papers

• Self-Explained Keywords Empower Large Language Models for Code Generation [5.236633572296712]
  Large language models (LLMs) have achieved impressive performance in code generation. Sek(textbfSelf-textbfExplained textbfKeywords) extracts and explains the key terms in the problem description with the LLM itself.
  arXiv  Detail & Related papers  (2024-10-21T12:52:03Z)
• CodecLM: Aligning Language Models with Tailored Synthetic Data [51.59223474427153]
  We introduce CodecLM, a framework for adaptively generating high-quality synthetic data for instruction-following abilities. We first encode seed instructions into metadata, which are concise keywords generated on-the-fly to capture the target instruction distribution. We also introduce Self-Rubrics and Contrastive Filtering during decoding to tailor data-efficient samples.
  arXiv  Detail & Related papers  (2024-04-08T21:15:36Z)
• CodeUltraFeedback: An LLM-as-a-Judge Dataset for Aligning Large Language Models to Coding Preferences [2.3749120526936465]
  We propose using the LLM-as-a-Judge methodology to evaluate the alignment of LLMs with coding preferences. CodeUltraFeedback consists of 10,000 coding instructions, each annotated with four responses generated from a diverse pool of 14 LLMs. In turn, we explore the usage of CodeUltraFeedback as feedback data to fine-tune and align CodeLlama-7B-Instruct using supervised fine-tuning (SFT) and reinforcement learning from AI feedback (RLAIF) with direct preference optimization (DPO)
  arXiv  Detail & Related papers  (2024-03-14T01:51:35Z)
• Direct Preference Optimization for Neural Machine Translation with Minimum Bayes Risk Decoding [15.309135455863753]
  We show how the recently developed Reinforcement Learning technique, Direct Preference Optimization (DPO), can fine-tune Multilingual Large Language Models without additional computation. Our method uses only a small monolingual fine-tuning set and yields significantly improved performance on multiple NMT test sets compared to MLLMs without DPO.
  arXiv  Detail & Related papers  (2023-11-14T18:43:51Z)
• Reflection-Tuning: Data Recycling Improves LLM Instruction-Tuning [79.32236399694077]
  Low-quality data in the training set are usually detrimental to instruction tuning. We propose a novel method, termed "reflection-tuning" This approach utilizes an oracle LLM to recycle the original training data by introspecting and enhancing the quality of instructions and responses in the data.
  arXiv  Detail & Related papers  (2023-10-18T05:13:47Z)
• Evaluating Large Language Models at Evaluating Instruction Following [54.49567482594617]
  We introduce a challenging meta-evaluation benchmark, LLMBar, designed to test the ability of an LLM evaluator in discerning instruction-following outputs. We discover that different evaluators exhibit distinct performance on LLMBar and even the highest-scoring ones have substantial room for improvement.
  arXiv  Detail & Related papers  (2023-10-11T16:38:11Z)
• LLMRec: Benchmarking Large Language Models on Recommendation Task [54.48899723591296]
  The application of Large Language Models (LLMs) in the recommendation domain has not been thoroughly investigated. We benchmark several popular off-the-shelf LLMs on five recommendation tasks, including rating prediction, sequential recommendation, direct recommendation, explanation generation, and review summarization. The benchmark results indicate that LLMs displayed only moderate proficiency in accuracy-based tasks such as sequential and direct recommendation.
  arXiv  Detail & Related papers  (2023-08-23T16:32:54Z)
• On Learning to Summarize with Large Language Models as References [101.79795027550959]
  Large language models (LLMs) are favored by human annotators over the original reference summaries in commonly used summarization datasets. We study an LLM-as-reference learning setting for smaller text summarization models to investigate whether their performance can be substantially improved.
  arXiv  Detail & Related papers  (2023-05-23T16:56:04Z)

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.