Large Language Models (LLMs) are often used as evaluators of text quality, known as LLM-as-a-Judge, which can outperform conventional automatic evaluation metrics that rely on reference texts. However, LLM evaluators tend to generate particular scores regardless of the context of the evaluated text, which is known as scoring bias.
arXiv:2608. 05726v1 Announce Type: cross Abstract: Large Language Models (LLMs) are often used as evaluators of text quality, known as LLM-as-a-Judge, which can outperform conventional automatic evaluation metrics that rely on reference texts.
The paper examines LLM-as-a-Judge systems used to assess AI-generated text, questioning the assumption that judgments are derived from reasoning over responses and rubrics. It finds that classifiers trained solely on rubric text can predict judge outputs, indicating that rubrics contain recoverable evaluative signals independent of the responses. Counterfactual experiments show judges often fail to adjust decisions when either the response or rubric criterion is reversed, raising doubts about the reliability of rubric-based LLM evaluation.
arXiv:2607. 05679v1 Announce Type: cross Abstract: Language models (LMs) exhibit problematic biases, such as stereotypes.
The study investigates how prior scores influence large language model (LLM) judgments in the LLM-as-a-Judge paradigm. By testing three prompt conditions—no metadata, revision framing, and anchored metadata containing prior scores—the authors find that prior scores systematically bias evaluations, shifting ratings toward those scores across 192,000 attempts. The bias also affects categorical decisions, blocking 48% of error corrections and flipping 10.18% of correct judgments, and is not mitigated by Chain-of-Thought or a warning, underscoring the need for careful context engineering.
Reference-based text evaluation metrics, which are widely used to assess natural language generation systems, score a candidate response by comparing it with a reference response. The reliability of an evaluation metric is usually judged by its statistical correlation with human ratings.
arXiv:2608. 01423v1 Announce Type: cross Abstract: Reference-based text evaluation metrics, which are widely used to assess natural language generation systems, score a candidate response by comparing it with a reference response.
arXiv:2609.24516v1 Announce Type: new Abstract: In recent years, large language models (LLMs) have emerged as a popular alternative for evaluation. Often referred to as LLMs as judges (LLJs), these s...
The paper introduces VOLM, a framework that quantifies how much original value a human adds to a document beyond what a language model could generate from a task description alone. Unlike existing tools that focus on stylistic detection, VOLM extracts content at varying granularities, reconstructs it with an LLM, and compares these reconstructions to those derived from the task description. Evaluations across news articles, ICLR peer reviews, and argumentative essays show that VOLM can distinguish human-authored texts from LLM-generated ones while remaining robust to content-preserving transformations.
The paper introduces BASSE, a multilingual meta‑evaluation dataset containing 2,040 human‑rated abstractive summaries produced manually or by five LLMs with four prompts. Annotators scored each summary on coherence, consistency, fluency, relevance, and 5W1H using a 5‑point Likert scale. Benchmarking shows proprietary LLM‑judge models best align with human judgments, followed by criteria‑specific automatic metrics, while open‑source judge LLMs perform poorly.
The paper introduces JudgeBiasBench, a benchmark that systematically quantifies judgment biases in large language model (LLM)-based judges across four dimensions and 12 bias types. It evaluates both generative and discriminative judges, revealing significant bias patterns that undermine reliability. The authors propose bias-aware training—reinforcement learning for generative judges and contrastive learning for discriminative judges—to reduce these biases while maintaining evaluation performance.
IDEAlign introduces a new protocol for evaluating the similarity of large language model (LLM) annotations to expert judgments. It uses pick‑the‑odd‑one‑out tasks to capture expert similarity and benchmarks various similarity methods—including text embeddings, topic models, and LLM-as-a-judge—against these human ratings. Applied to educational datasets, the study finds that most metrics miss nuanced expert dimensions, with LLM-as-a-judge performing best yet still insufficient for full expert alignment.
Rubric-based evaluation is a promising approach for assessing open-ended outputs from LLM-based research agents, particularly in paper reproduction, where direct paper-to-repository comparison is prone to hallucination. However, constructing paper-specific rubrics requires substantial expert effort, limiting the scalability of benchmarks such as PaperBench.