Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena

Abstract

Evaluating large language model (LLM) based chat assistants is challenging due to their broad capabilities and the inadequacy of existing benchmarks in measuring human preferences. To address this, we explore using strong LLMs as judges to evaluate these models on more open-ended questions. We examine the usage and limitations of LLM-as-a-judge, including position, verbosity, and self-enhancement biases, as well as limited reasoning ability, and propose solutions to mitigate some of them. We then verify the agreement between LLM judges and human preferences by introducing two benchmarks: MT-bench, a multi-turn question set; and Chatbot Arena, a crowdsourced battle platform. Our results reveal that strong LLM judges like GPT-4 can match both controlled and crowdsourced human preferences well, achieving over 80% agreement, the same level of agreement between humans. Hence, LLM-as-a-judge is a scalable and explainable way to approximate human preferences, which are otherwise very expensive to obtain. Additionally, we show our benchmark and traditional benchmarks complement each other by evaluating several variants of LLaMA and Vicuna. The MT-bench questions, 3K expert votes, and 30K conversations with human preferences are publicly available at https://github.com/lm-sys/FastChat/tree/main/fastchat/llm_judge.

NeurIPS 2023 Datasets and Benchmarks Track

Cited by in corpus (39)

• Adapted Large Language Models Can Outperform Medical Experts in Clinical Text Summarization
• Evaluation of Retrieval-Augmented Generation: A Survey
• ChatEDA: A Large Language Model Powered Autonomous Agent for EDA
• Taking the Next Step with Generative Artificial Intelligence: The Transformative Role of Multimodal Large Language Models in Science Education
• MasterKey: Automated Jailbreak Across Multiple Large Language Model Chatbots
• Theory of Mind for Multi-Agent Collaboration via Large Language Models
• Evaluation and Benchmarking of LLM Agents: A Survey
• Evaluation of LLM Chatbots for OSINT-based Cyber Threat Awareness
• On the Prospects of Incorporating Large Language Models (LLMs) in Automated Planning and Scheduling (APS)
• A Survey on Occupancy Perception for Autonomous Driving: The Information Fusion Perspective
• Could ChatGPT get an Engineering Degree? Evaluating Higher Education Vulnerability to AI Assistants
• GeoCode-GPT: A Large Language Model for Geospatial Code Generation Tasks
• Automated Review Generation Method Based on Large Language Models
• Towards Reliable Generative AI-Driven Scaffolding: Reducing Hallucinations and Enhancing Quality in Self-Regulated Learning Support
• CACTUS: Chemistry Agent Connecting Tool-Usage to Science
• Evaluating Large Language Models in Process Mining: Capabilities, Benchmarks, and Evaluation Strategies
• Query Performance Prediction using Relevance Judgments Generated by Large Language Models
• Large Language Models, scientific knowledge and factuality: A framework to streamline human expert evaluation
• Benchmarking Uncertainty Quantification Methods for Large Language Models with LM-Polygraph
• Foundation Models and Transformers for Anomaly Detection: A Survey
• Large Language Models for Scholarly Ontology Generation: An Extensive Analysis in the Engineering Field
• Speed and Conversational Large Language Models: Not All Is About Tokens per Second
• Benchmarking LLM-based Relevance Judgment Methods
• Democratizing LLMs: An Exploration of Cost-Performance Trade-offs in Self-Refined Open-Source Models
• Revisiting Block-based Quantisation: What is Important for Sub-8-bit LLM Inference?
• Enhancing chest X-ray datasets with privacy-preserving large language models and multi-type annotations: a data-driven approach for improved classification
• The Strong Pull of Prior Knowledge in Large Language Models and Its Impact on Emotion Recognition
• SynthTRIPs: A Knowledge-Grounded Framework for Benchmark Query Generation for Personalized Tourism Recommenders
• EvidenceMap: Learning Evidence Analysis to Unleash the Power of Small Language Models for Biomedical Question Answering
• Advancing Risk and Quality Assurance: A RAG Chatbot for Improved Regulatory Compliance
• FAITH: A Framework for Assessing Intrinsic Tabular Hallucinations in Finance
• Don't Get Too Excited -- Eliciting Emotions in LLMs
• Can Large Language Models Assess Serendipity in Recommender Systems?
• CPR: Mitigating Large Language Model Hallucinations with Curative Prompt Refinement
• Toward a benchmark for CTR prediction in online advertising: datasets, evaluation protocols and perspectives
• Towards AI epidemiology: a measurement standardisation framework for prospective risk detection
• A Hierarchical Framework for Measuring Scientific Paper Innovation via Large Language Models
• Linguacodus: A Synergistic Framework for Transformative Code Generation in Machine Learning Pipelines
• Reliable Annotations with Less Effort: Evaluating LLM-Human Collaboration in Search Clarifications