Time To Impeach LLM-as-a-Judge: Programs are the Future of Evaluation

• URL: http://arxiv.org/abs/2506.10403v1
• Date: Thu, 12 Jun 2025 06:53:22 GMT
• Title: Time To Impeach LLM-as-a-Judge: Programs are the Future of Evaluation
• Authors: Tzu-Heng Huang, Harit Vishwakarma, Frederic Sala,
• Abstract summary: Large language models (LLMs) are widely used to evaluate the quality of LLM generations and responses.<n>We introduce PAJAMA, a new alternative that uses LLMs to synthesize executable judging programs instead of directly scoring responses.<n>These synthesized programs can be stored and run locally, costing orders of magnitude less while providing interpretable, and auditable judging logic.
• Score: 16.017941636877193
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models (LLMs) are widely used to evaluate the quality of LLM generations and responses, but this leads to significant challenges: high API costs, uncertain reliability, inflexible pipelines, and inherent biases. To address these, we introduce PAJAMA (Program-As-a-Judge for Automated Model Assessment), a new alternative that uses LLMs to synthesize executable judging programs instead of directly scoring responses. These synthesized programs can be stored and run locally, costing orders of magnitude less while providing interpretable, and auditable judging logic that can be easily adapted. Program-based judges mitigate biases, improving judgment consistency by 15.83% and reducing biased responses by 23.7% on average compared to a Qwen2.5-14B-based LLM-as-a-judge. When program judgments are distilled into a model, PAJAMA outperforms LLM-as-a-judge on the challenging CHAT-HARD subset of RewardBench, outperforming metrics by 2.19% on Prometheus and 8.67% on the JudgeLM dataset, all at three orders of magnitude lower cost.

Related papers

• CodeJudgeBench: Benchmarking LLM-as-a-Judge for Coding Tasks [63.562924932512765]
  Large Language Models (LLMs) have advanced the state-of-the-art in various coding tasks.<n>LLMs can also serve as judges, assessing and comparing the quality of responses generated by other models.
  arXiv  Detail & Related papers  (2025-07-14T17:56:29Z)
• Reverse Engineering Human Preferences with Reinforcement Learning [14.508050809497847]
  Large Language Models (LLMs) are routinely evaluated by other LLMs trained to predict human preferences.<n>Previous work shows that the answers generated by a candidate-LLM can be edited post hoc to maximise the score assigned to them by a judge-LLM.<n>We adopt a different approach and use the signal provided by judge-LLMs as a reward to adversarially tune models.
  arXiv  Detail & Related papers  (2025-05-21T17:48:16Z)
• Leveraging LLMs as Meta-Judges: A Multi-Agent Framework for Evaluating LLM Judgments [6.270885758858811]
  Large language models (LLMs) are being widely applied across various fields, but as tasks become more complex, evaluating their responses is increasingly challenging.<n>We propose a three-stage meta-judge selection pipeline: 1) developing a comprehensive rubric with GPT-4 and human experts, 2) using three advanced LLM agents to score judgments, and 3) applying a threshold to filter out low-scoring judgments.<n> Experimental results on the JudgeBench dataset show about 15.55% improvement compared to raw judgments and about 8.37% improvement over the single-agent baseline.
  arXiv  Detail & Related papers  (2025-04-23T20:32:12Z)
• 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.<n>It evaluates judge performance in three domains (math reasoning, code generation, and instruction following) under three task settings.<n>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)
• LLM-as-a-Judge: Reassessing the Performance of LLMs in Extractive QA [27.86922657261678]
  We reassess the performance of QA models using LLM-as-a-judge across four reading comprehension QA datasets.<n>Our results show that LLM-as-a-judge is highly correlated with human judgments and can replace traditional EM/F1 metrics.
  arXiv  Detail & Related papers  (2025-04-16T11:08:46Z)
• A Judge-free LLM Open-ended Generation Benchmark Based on the Distributional Hypothesis [1.5802986215292303]
  We propose a novel benchmark that evaluates large language models (LLMs) using n-gram statistics and rules.<n>Using 50 question and reference answer sets, we introduce three new metrics based on n-grams and rules: Fluency, Truthfulness, and Helpfulness.<n>Our benchmark strongly correlates with GPT-4o-based evaluations while requiring significantly fewer computational resources.
  arXiv  Detail & Related papers  (2025-02-13T13:30:54Z)
• Tuning LLM Judge Design Decisions for 1/1000 of the Cost [42.06346155380305]
  Large Language Models (LLMs) often require costly human annotations.<n>To address this, LLM-based judges have been proposed, which compare the outputs of two LLMs.<n>While several approaches have been proposed, many confounding factors are present between different papers.
  arXiv  Detail & Related papers  (2025-01-24T17:01:14Z)
• 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.<n>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)
• MQM-APE: Toward High-Quality Error Annotation Predictors with Automatic Post-Editing in LLM Translation Evaluators [53.91199933655421]
  Large Language Models (LLMs) have shown significant potential as judges for Machine Translation (MT) quality assessment.<n>We introduce a universal and training-free framework, $textbfMQM-APE, based on the idea of filtering out non-impactful errors.<n>Experiments show that our approach consistently improves both the reliability and quality of error spans against GEMBA-MQM.
  arXiv  Detail & Related papers  (2024-09-22T06:43:40Z)
• MixEval: Deriving Wisdom of the Crowd from LLM Benchmark Mixtures [57.886592207948844]
  We propose MixEval, a new paradigm for establishing efficient, gold-standard evaluation by strategically mixing off-the-shelf benchmarks. It bridges (1) comprehensive and well-distributed real-world user queries and (2) efficient and fairly-graded ground-truth-based benchmarks, by matching queries mined from the web with similar queries from existing benchmarks.
  arXiv  Detail & Related papers  (2024-06-03T05:47:05Z)
• Self-Evaluation Improves Selective Generation in Large Language Models [54.003992911447696]
  We reformulate open-ended generation tasks into token-level prediction tasks. We instruct an LLM to self-evaluate its answers. We benchmark a range of scoring methods based on self-evaluation.
  arXiv  Detail & Related papers  (2023-12-14T19:09:22Z)

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.