Quantifying and Mitigating Self-Preference Bias of LLM Judges

arXiv:2604.22891v1 Announce Type: cross Abstract: LLM-as-a-Judge has become a dominant approach in automated evaluation systems, playing critical roles in model alignment, leaderboard construction, quality control, and so on. However, the scalability and trustworthiness of this approach can be substantially distorted by Self-Preference Bias (SPB), which is a directional evaluative deviation in which LLMs systematically favor or disfavor their own generated outputs during evaluation. Existing measurements rely on costly human annotations and conflate generative capability with evaluative stance, and thus are impractical for large-scale deployment in real-world systems. To address this issue, we introduce a fully automated framework to quantifying and mitigating SPB, which constructs equal-quality pairs of responses with negligible quality differences, enabling statistical disentanglement of discriminability from bias propensity without human gold standards. Empirical analysis across 20 mainstream LLMs reveals that advanced capabilities are often uncorrelated, or even negatively correlated, with low SPB. To mitigate this bias, we propose a structured multi-dimensional evaluation strategy grounded in cognitive load decomposition, which reduces SPB by 31.5% on average.

Related

• Peer Identity Bias in Multi-Agent LLM Evaluation: An Empirical Study Using the TRUST Democratic Discourse Analysis Pipeline
• Self-Preference Bias in Rubric-Based Evaluation of Large Language Models
• Do LLM Decoders Listen Fairly? Benchmarking How Language Model Priors Shape Bias in Speech Recognition
• Measuring and Exploiting Contextual Bias in LLM-Assisted Security Code Review
• Detection Is Cheap, Routing Is Learned: Why Refusal-Based Alignment Evaluation Fails

Source: arXiv cs.AI | 2026-04-28