From Calculation to Adjudication: Examining LLM judges on Mathematical Reasoning Tasks

• URL: http://arxiv.org/abs/2409.04168v2
• Date: Mon, 12 May 2025 19:41:57 GMT
• Title: From Calculation to Adjudication: Examining LLM judges on Mathematical Reasoning Tasks
• Authors: Andreas Stephan, Dawei Zhu, Matthias Aßenmacher, Xiaoyu Shen, Benjamin Roth,
• Abstract summary: We study large language models (LLMs) on mathematical reasoning tasks.<n>Our analysis uncovers a strong correlation between judgment performance and the candidate model task performance.<n>As a consequence, we test whether we can predict the behavior of LLM judges using simple features such as part-of-speech tags.
• Score: 11.01213914485374
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: To reduce the need for human annotations, large language models (LLMs) have been proposed as judges of the quality of other candidate models. The performance of LLM judges is typically evaluated by measuring the correlation with human judgments on generative tasks such as summarization or machine translation. In contrast, we study LLM judges on mathematical reasoning tasks. These tasks require multi-step reasoning, and the correctness of their solutions is verifiable, enabling a more objective evaluation. We perform a detailed performance analysis and find that easy samples are easy to judge, and difficult samples are difficult to judge. Our analysis uncovers a strong correlation between judgment performance and the candidate model task performance, indicating that judges tend to favor higher-quality models even if their answer is incorrect. As a consequence, we test whether we can predict the behavior of LLM judges using simple features such as part-of-speech tags and find that we can correctly predict 70%-75% of judgments. We conclude this study by analyzing practical use cases, showing that LLM judges consistently detect the on-average better model but largely fail if we use them to improve task performance.

Related papers

• Quantitative LLM Judges [48.676042957523045]
  We propose quantitative LLM judges, which align evaluation scores of existing LLM judges to human scores in a given domain.<n>The models are trained to improve the score of the original judge by using the judge's textual evaluation and score.<n>Our experiments show that quantitative judges can effectively improve the predictive power of existing judges through post-hoc modeling.
  arXiv  Detail & Related papers  (2025-06-03T14:44:23Z)
• MR. Judge: Multimodal Reasoner as a Judge [23.787019892923784]
  We propose Multimodal Reasoner as a Judge (MR. Judge) as a paradigm for empowering general-purpose MLLMs judges with strong reasoning capabilities.<n>Instead of directly assigning scores for each response, we formulate the judgement process as a reasoning-inspired multiple-choice problem.<n>This reasoning process not only improves the interpretibility of the judgement, but also greatly enhances the performance of MLLM judges.
  arXiv  Detail & Related papers  (2025-05-19T17:37:39Z)
• Evaluating Judges as Evaluators: The JETTS Benchmark of LLM-as-Judges as Test-Time Scaling Evaluators [66.83088028268318]
  This paper introduces the Judge Evaluation for Test-Time Scaling benchmark. It evaluates judge performance in three domains (math reasoning, code generation, and instruction following) under three task settings. Our benchmark shows that while judges are competitive with outcome reward models in reranking, they are consistently worse than process reward models in beam search procedures.
  arXiv  Detail & Related papers  (2025-04-21T17:33:23Z)
• JudgeLRM: Large Reasoning Models as a Judge [65.14085339820795]
  We investigate whether Large Language Models (LLMs) judges truly benefit from enhanced reasoning capabilities.<n>We introduce JudgeLRM, a family of judgment-oriented LLMs trained using reinforcement learning (RL) with judge-wise, outcome-driven rewards.
  arXiv  Detail & Related papers  (2025-03-31T02:18:51Z)
• Verdict: A Library for Scaling Judge-Time Compute [4.962699700524792]
  Verdict is a library for scaling judge-time compute to enhance the accuracy, reliability, and interpretability of automated evaluators. Verdict judges achieve state-of-the-art (SOTA) or near-SOTA performance, surpassing orders-of-magnitude larger fine-tuned judges.
  arXiv  Detail & Related papers  (2025-02-25T09:26:44Z)
• JuStRank: Benchmarking LLM Judges for System Ranking [7.507819077549208]
  We conduct the first large-scale study of LLM judges as system rankers. System scores are generated by aggregating judgment scores over multiple system outputs. Our analysis provides a fine-grained characterization of judge behavior, including their decisiveness and bias.
  arXiv  Detail & Related papers  (2024-12-12T18:51:13Z)
• JudgeRank: Leveraging Large Language Models for Reasoning-Intensive Reranking [81.88787401178378]
  We introduce JudgeRank, a novel agentic reranker that emulates human cognitive processes when assessing document relevance. We evaluate JudgeRank on the reasoning-intensive BRIGHT benchmark, demonstrating substantial performance improvements over first-stage retrieval methods. In addition, JudgeRank performs on par with fine-tuned state-of-the-art rerankers on the popular BEIR benchmark, validating its zero-shot generalization capability.
  arXiv  Detail & Related papers  (2024-10-31T18:43:12Z)
• JudgeBench: A Benchmark for Evaluating LLM-based Judges [61.048125269475854]
  JudgeBench is a benchmark for evaluating LLM-based judges on challenging response pairs spanning knowledge, reasoning, math, and coding. Our comprehensive evaluation on a collection of prompted judges, fine-tuned judges, multi-agent judges, and reward models shows that JudgeBench poses a significantly greater challenge than previous benchmarks.
  arXiv  Detail & Related papers  (2024-10-16T17:58:19Z)
• Evaluating the Evaluator: Measuring LLMs' Adherence to Task Evaluation Instructions [18.93335792080899]
  We investigate how much influence prompting the LLMs-as-a-judge has on the alignment of AI judgements to human judgements. We aggregate a taxonomy of quality criteria commonly used across state-of-the-art evaluations with LLMs and provide this as a rigorous benchmark of models as judges.
  arXiv  Detail & Related papers  (2024-08-16T14:49:35Z)
• On scalable oversight with weak LLMs judging strong LLMs [67.8628575615614]
  We study debate, where two AI's compete to convince a judge; consultancy, where a single AI tries to convince a judge that asks questions. We use large language models (LLMs) as both AI agents and as stand-ins for human judges, taking the judge models to be weaker than agent models.
  arXiv  Detail & Related papers  (2024-07-05T16:29:15Z)
• Judging the Judges: Evaluating Alignment and Vulnerabilities in LLMs-as-Judges [6.609843448260634]
  The LLM-as-a-judge paradigm is rapidly gaining traction as an approach to evaluating large language models. This paper focuses on a clean scenario in which inter-human agreement is high. We identify vulnerabilities in judge models, such as their sensitivity to prompt complexity and length, and a tendency toward leniency.
  arXiv  Detail & Related papers  (2024-06-18T13:49:54Z)
• Language Model Council: Democratically Benchmarking Foundation Models on Highly Subjective Tasks [3.58262772907022]
  We introduce the Language Model Council (LMC), where a group of LLMs collaborate to create tests, respond to them, and evaluate each other's responses to produce a ranking in a democratic fashion.<n>In a detailed case study on emotional intelligence, we deploy a council of 20 recent LLMs to rank each other on open-ended responses to interpersonal conflicts.<n>Our results show that the LMC produces rankings that are more separable and more robust, and through a user study, we show that they are more consistent with human evaluations than any individual LLM judge.
  arXiv  Detail & Related papers  (2024-06-12T19:05:43Z)
• Aligning Large Language Models by On-Policy Self-Judgment [49.31895979525054]
  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.
  arXiv  Detail & Related papers  (2024-02-17T11:25:26Z)
• Exploring the Reliability of Large Language Models as Customized Evaluators for Diverse NLP Tasks [65.69651759036535]
  We analyze whether large language models (LLMs) can serve as reliable alternatives to humans.<n>This paper explores both conventional tasks (e.g., story generation) and alignment tasks (e.g., math reasoning)<n>We find that LLM evaluators can generate unnecessary criteria or omit crucial criteria, resulting in a slight deviation from the experts.
  arXiv  Detail & Related papers  (2023-10-30T17:04:35Z)
• JudgeLM: Fine-tuned Large Language Models are Scalable Judges [54.007823006976516]
  We propose to fine-tune Large Language Models (LLMs) as scalable judges (JudgeLM) to evaluate LLMs efficiently and effectively in open-ended benchmarks. We first propose a comprehensive, large-scale, high-quality dataset containing task seeds, LLMs-generated answers, and GPT-4-generated judgments for fine-tuning high-performance judges. We then analyze the key biases in fine-tuning LLM as a judge and consider them as position bias, knowledge bias, and format bias.
  arXiv  Detail & Related papers  (2023-10-26T17:48:58Z)
• Evaluating the Performance of Large Language Models on GAOKAO Benchmark [53.663757126289795]
  This paper introduces GAOKAO-Bench, an intuitive benchmark that employs questions from the Chinese GAOKAO examination as test samples. With human evaluation, we obtain the converted total score of LLMs, including GPT-4, ChatGPT and ERNIE-Bot. We also use LLMs to grade the subjective questions, and find that model scores achieve a moderate level of consistency with human scores.
  arXiv  Detail & Related papers  (2023-05-21T14:39:28Z)
• Perspectives on Large Language Models for Relevance Judgment [56.935731584323996]
  Large language models (LLMs) claim that they can assist with relevance judgments. It is not clear whether automated judgments can reliably be used in evaluations of retrieval systems.
  arXiv  Detail & Related papers  (2023-04-13T13:08:38Z)
• Benchmarking Large Language Models for News Summarization [79.37850439866938]
  Large language models (LLMs) have shown promise for automatic summarization but the reasons behind their successes are poorly understood. We find instruction tuning, and not model size, is the key to the LLM's zero-shot summarization capability.
  arXiv  Detail & Related papers  (2023-01-31T18:46:19Z)

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.