MLLM-as-a-Judge Exhibits Model Preference Bias

• URL: http://arxiv.org/abs/2604.11589v1
• Date: Mon, 13 Apr 2026 15:04:40 GMT
• Title: MLLM-as-a-Judge Exhibits Model Preference Bias
• Abstract summary: It remains unclear to what extent MLLM-as-a-Judge methods favor or disfavor text generated by specific MLLMs.<n>Using 1.29M caption-score pairs collected from 12 MLLMs, we found that representative MLLMs tend to exhibit self-preference bias.<n>Our results demonstrated that Pomms effectively mitigated the model-specific preference bias while maintaining performance.
• Score: 3.369875207401089
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Automatic evaluation using multimodal large language models (MLLMs), commonly referred to as MLLM-as-a-Judge, has been widely used to measure model performance. If such MLLM-as-a-Judge methods were biased, they could distort model comparisons and benchmark-driven scientific progress. However, it remains unclear to what extent MLLM-as-a-Judge methods favor or disfavor text generated by specific MLLMs. In this study, we propose Philautia-Eval to investigate such model-specific preference bias. Philautia-Eval quantifies the degree of the bias by disentangling preference tendencies from differences in generation quality. Using 1.29M caption-score pairs collected from 12 MLLMs, we found that representative MLLMs tend to exhibit self-preference bias. Moreover, experimental results indicate mutual preference bias within particular model families, which is potentially driven by reused connectors and overlapping instruction-tuning resources. Finally, we introduce a simple ensemble of MLLMs, Pomms. Our results demonstrated that Pomms effectively mitigated the model-specific preference bias while maintaining performance.

Related papers

• Beyond Accuracy: LLM Variability in Evidence Screening for Software Engineering SLRs [0.29737220923371155]
  Large Language Models (LLMs) are used for study screening in systematic literature reviews.<n>Despite rapid uptake, there is limited evidence on how such models behave during the study screening phase.<n>This study investigated the impact of input metadata (abstract, title, keywords) on LLM performance.
  arXiv  Detail & Related papers  (2026-04-29T08:55:16Z)
• Blind to the Human Touch: Overlap Bias in LLM-Based Summary Evaluation [89.52571224447111]
  Large language model (LLM) judges have often been used alongside traditional, algorithm-based metrics for tasks like summarization.<n>We provide an LLM judge bias analysis as a function of overlap with human-written responses in the domain of summarization.
  arXiv  Detail & Related papers  (2026-02-07T19:39:28Z)
• Nonparametric LLM Evaluation from Preference Data [86.96268870461472]
  We propose a nonparametric statistical framework, DMLEval, for comparing and ranking large language models (LLMs) from preference data.<n>Our framework provides practitioners with powerful, state-of-the-art methods for comparing or ranking LLMs.
  arXiv  Detail & Related papers  (2026-01-29T15:00:07Z)
• Play Favorites: A Statistical Method to Measure Self-Bias in LLM-as-a-Judge [17.40713507922006]
  Large language models (LLMs) can serve as judges that offer rapid and reliable assessments of other outputs.<n>LLMs may systematically assign overly favorable ratings to their own outputs, a phenomenon known as self-bias.<n>We present a statistical framework that explicitly formalizes assumptions under which self-bias can be identified and estimated.
  arXiv  Detail & Related papers  (2025-08-08T21:22:12Z)
• Relative Bias: A Comparative Framework for Quantifying Bias in LLMs [29.112649816695203]
  Relative Bias is a method designed to assess how an LLM's behavior deviates from other LLMs within a specified target domain.<n>We introduce two complementary methodologies: (1) Embedding Transformation analysis, which captures relative bias patterns through sentence representations over the embedding space, and (2) LLM-as-a-Judge, which employs a language model to evaluate outputs comparatively.<n>Applying our framework to several case studies on bias and alignment scenarios following by statistical tests for validation, we find strong alignment between the two scoring methods.
  arXiv  Detail & Related papers  (2025-05-22T01:59:54Z)
• Preference Leakage: A Contamination Problem in LLM-as-a-judge [69.96778498636071]
  Large Language Models (LLMs) as judges and LLM-based data synthesis have emerged as two fundamental LLM-driven data annotation methods.<n>In this work, we expose preference leakage, a contamination problem in LLM-as-a-judge caused by the relatedness between the synthetic data generators and LLM-based evaluators.
  arXiv  Detail & Related papers  (2025-02-03T17:13:03Z)
• Self-Preference Bias in LLM-as-a-Judge [13.880151307013321]
  We introduce a novel metric to measure the self-preference bias in large language models (LLMs)<n>Our results show GPT-4 exhibits a significant degree of self-preference bias.<n>This suggests that the essence of the bias lies in perplexity and that the self-preference bias exists because LLMs prefer texts more familiar to them.
  arXiv  Detail & Related papers  (2024-10-29T07:42:18Z)
• Justice or Prejudice? Quantifying Biases in LLM-as-a-Judge [84.34545223897578]
  Despite their excellence in many domains, potential issues are under-explored, undermining their reliability and the scope of their utility. We identify 12 key potential biases and propose a new automated bias quantification framework-CALM- which quantifies and analyzes each type of bias in LLM-as-a-Judge. Our work highlights the need for stakeholders to address these issues and remind users to exercise caution in LLM-as-a-Judge applications.
  arXiv  Detail & Related papers  (2024-10-03T17:53:30Z)
• A Gradient Analysis Framework for Rewarding Good and Penalizing Bad Examples in Language Models [63.949883238901414]
  We present a unique angle of gradient analysis of loss functions that simultaneously reward good examples and penalize bad ones in LMs. We find that ExMATE serves as a superior surrogate for MLE, and that combining DPO with ExMATE instead of MLE further enhances both the statistical (5-7%) and generative (+18% win rate) performance.
  arXiv  Detail & Related papers  (2024-08-29T17:46:18Z)
• Self-Exploring Language Models: Active Preference Elicitation for Online Alignment [88.56809269990625]
  We propose a bilevel objective optimistically biased towards potentially high-reward responses to actively explore out-of-distribution regions. Our experimental results demonstrate that when fine-tuned on Zephyr-7B-SFT and Llama-3-8B-Instruct models, Self-Exploring Language Models (SELM) significantly boosts the performance on instruction-following benchmarks.
  arXiv  Detail & Related papers  (2024-05-29T17:59:07Z)
• Large Language Model (LLM) Bias Index -- LLMBI [0.0]
  The Large Language Model Bias Index (LLMBI) is a pioneering approach designed to quantify and address biases inherent in large language models (LLMs) We formulated LLMBI using a composite scoring system incorporating multiple dimensions of bias, including but not limited to age, gender, and racial biases. Our empirical analysis, conducted using responses from OpenAI's API, employs advanced sentiment analysis as a representative method for bias detection.
  arXiv  Detail & Related papers  (2023-12-22T15:38:13Z)

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.