Evaluating Scoring Bias in LLM-as-a-Judge
This addresses a critical issue for researchers and practitioners using LLMs as evaluators in fields like NLP and preference learning, though it is incremental as it builds on existing benchmarks.
The paper tackles the problem of scoring bias in LLM-as-a-Judge systems, where biases in scoring-based evaluations affect fairness and reliability, and provides a framework to evaluate this bias, finding that existing judge models' scoring stability is disrupted by such biases.
The remarkable performance of Large Language Models (LLMs) gives rise to``LLM-as-a-Judge'', where LLMs are employed as evaluators for complex tasks. Moreover, it has been widely adopted across fields such as Natural Language Processing (NLP), preference learning, and various specific domains. However, there are various biases within LLM-as-a-Judge, which adversely affect the fairness and reliability of judgments. Current research on evaluating or mitigating bias in LLM-as-a-Judge predominantly focuses on comparison-based evaluations, while systematic investigations into bias in scoring-based evaluations remain limited. Therefore, we define scoring bias in LLM-as-a-Judge as the scores differ when scoring judge models are bias-related perturbed, and provide a well-designed framework to comprehensively evaluate scoring bias. We augment existing LLM-as-a-Judge benchmarks through data synthesis to construct our evaluation dataset and design multi-faceted evaluation metrics. Our experimental results demonstrate that the scoring stability of existing judge models is disrupted by scoring biases. Further exploratory experiments and discussions provide valuable insights into the design of scoring prompt templates and the mitigation of scoring biases on aspects such as score rubrics, score IDs, and reference answer selection.