paper-with-me

홈 › Papers

When LLMs Benchmark Themselves: Deconstructing Self-Bias in Automated Evaluation

2025-09-30 · Wenda Xu, Sweta Agrawal, Vilém Zouhar, Markus Freitag, Daniel Deutsch arxiv

As LLMs rapidly saturate existing benchmarks, automated benchmark creation using LLMs (LLM-as-a-benchmark) -- where a model generates test inputs (LLM-as-a-testset) and evaluates outputs (LLM-as-an-evaluator) -- has gained traction as a cheap alternative to human curation. We show that this paradigm has a fundamental problem: LLM-generated benchmarks systematically favor the model that created them. Using machine translation as our primary testbed, we find that self-bias arises from two additive sources, LLM-as-a-testset and LLM-as-an-evaluator, and their combination amplifies the effect. Crucially, even when test data is generated with explicit diversity controls, each model's implicit stylistic tendencies produce homogeneous, model-specific outputs that inflate its own scores. Increasing source text diversity, using our proposed diversity metric, partially mitigates this bias. Self-bias is strong enough to cause each model to rank itself first, overriding the peer-consensus ordering. We confirm that the phenomenon extends to open-ended generation on the Chatbot Arena task.

📄 PDF Abstract BibTeX arXiv:2509.26600

Code (0)

등록된 구현이 없습니다.

Tasks

Machine Translation

Similar Papers 제목 키워드 기반

LLM Evaluators Recognize and Favor Their Own Generations

2024-04-15 · Arjun Panickssery, Samuel R. Bowman, Shi Feng

Self-evaluation using large language models (LLMs) has proven valuable not only in benchmarking but also methods like reward modeling, constitutional AI, and self-refinement. But new biases are introduced due to the same…

Benchmarking

Can LLMs Correct Themselves? A Benchmark of Self-Correction in LLMs

2025-10-17 · Guiyao Tie, Zenghui Yuan, Zeli Zhao, Chaoran Hu 외 arxiv

Self-correction of large language models (LLMs) emerges as a critical component for enhancing their reasoning performance. Although various self-correction methods have been proposed, a comprehensive evaluation of these …

Mathematical ReasoningCode Generation

LLM Maybe LongLM: Self-Extend LLM Context Window Without Tuning

2024-01-02 · Hongye Jin, Xiaotian Han, Jingfeng Yang, Zhimeng Jiang 외

It is well known that LLMs cannot generalize well to long contexts whose lengths are larger than the training sequence length. This poses challenges when employing LLMs for processing long input sequences during inferenc…

The Self-Execution Benchmark: Measuring LLMs' Attempts to Overcome Their Lack of Self-Execution

2025-08-17 · Elon Ezra, Ariel Weizman, Amos Azaria arxiv

Large language models (LLMs) are commonly evaluated on tasks that test their knowledge or reasoning abilities. In this paper, we explore a different type of evaluation: whether an LLM can predict aspects of its own respo…

Efficient Self-Improvement in Multimodal Large Language Models: A Model-Level Judge-Free Approach

2024-11-26 · Shijian Deng, Wentian Zhao, Yu-Jhe Li, Kun Wan 외

Self-improvement in multimodal large language models (MLLMs) is crucial for enhancing their reliability and robustness. However, current methods often rely heavily on MLLMs themselves as judges, leading to high computati…

Hallucination