# Evaluating LLM-as-a-Judge Beyond Score Alignment: A Psychometric Analysis of Residual Judging Difficulty

arXiv: https://arxiv.org/abs/2610.02877 · HTML: https://mlsanity.com/paper/2610.02877/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 48 references that are in the map.
- First posted on arXiv 2 Oct 2026.

## Builds on (most cited first, 24 of 24)

- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [A Unified Approach to Interpreting Model Predictions](https://mlsanity.com/paper/1705.07874/) (2017)
- [G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment](https://mlsanity.com/paper/2303.16634/) (2023)
- [A Survey on LLM-as-a-Judge](https://mlsanity.com/paper/2411.15594/) (2024)
- [SummEval: Re-evaluating Summarization Evaluation](https://mlsanity.com/paper/2007.12626/) (2020)
- [Evaluating the Factual Consistency of Abstractive Text Summarization](https://mlsanity.com/paper/1910.12840/) (2019)
- [From Generation to Judgment: Opportunities and Challenges of LLM-as-a-judge](https://mlsanity.com/paper/2411.16594/) (2024)
- [Asking and Answering Questions to Evaluate the Factual Consistency of Summaries](https://mlsanity.com/paper/2004.04228/) (2020)
- [MLLM-as-a-Judge: Assessing Multimodal LLM-as-a-Judge with Vision-Language Benchmark](https://mlsanity.com/paper/2402.04788/) (2024)
- [Justice or Prejudice? Quantifying Biases in LLM-as-a-Judge](https://mlsanity.com/paper/2410.02736/) (2024)
- [PandaLM: An Automatic Evaluation Benchmark for LLM Instruction Tuning Optimization](https://mlsanity.com/paper/2306.05087/) (2023)
- [OpenChat: Advancing Open-source Language Models with Mixed-Quality Data](https://mlsanity.com/paper/2309.11235/) (2023)
- [LLMs instead of Human Judges? A Large Scale Empirical Study across 20 NLP Evaluation Tasks](https://mlsanity.com/paper/2406.18403/) (2024)
- [Large Language Models for Data Annotation and Synthesis: A Survey](https://mlsanity.com/paper/2402.13446/) (2024)
- [SOTOPIA-π: Interactive Learning of Socially Intelligent Language Agents](https://mlsanity.com/paper/2403.08715/) (2024)
- [GO FIGURE: A Meta Evaluation of Factuality in Summarization](https://mlsanity.com/paper/2010.12834/) (2020)
- [(A)I Am Not a Lawyer, But...: Engaging Legal Experts towards Responsible LLM Policies for Legal Advice](https://mlsanity.com/paper/2402.01864/) (2024)
- [Large Language Model Psychometrics: A Systematic Review of Evaluation, Validation, and Enhancement](https://mlsanity.com/paper/2505.08245/) (2025)
- [Human-Centered Design Recommendations for LLM-as-a-Judge](https://mlsanity.com/paper/2407.03479/) (2024)
- [Improve LLM-as-a-Judge Ability as a General Ability](https://mlsanity.com/paper/2502.11689/) (2025)
- [DiscoScore: Evaluating Text Generation with BERT and Discourse Coherence](https://mlsanity.com/paper/2201.11176/) (2022)
- [Role of Human-AI Interaction in Selective Prediction](https://mlsanity.com/paper/2112.06751/) (2021)
- [Diagnosing the Reliability of LLM-as-a-Judge via Item Response Theory](https://mlsanity.com/paper/2602.00521/) (2026)
- [Comparing Human and AI Rater Effects Using the Many-Facet Rasch Model](https://mlsanity.com/paper/2505.18486/) (2025)

## Abstract (arXiv)

Large language models (LLMs) are widely used as automatic judges, with validity typically assessed via alignment with human scores. However, aggregate agreement fails to reveal whether humans and LLMs find the same evaluation cases difficult. In this paper, we study this problem in summarization evaluation from a psychometric perspective. We fit Many-Facet Rasch Models separately to human and LLM ratings to decompose scores into latent summary quality, rater severity, dimension severity, and rating-scale thresholds. Building on this decomposition, we define residual hardness as a model-adjusted measure of judging difficulty and compare whether human and LLM judges share the same hardness structure. Across 17 open-weight LLM judges on SummEval, we find that moderate alignment in latent summary quality does not imply alignment in residual hardness. Human and LLM judges differ in which summary--dimension units remain difficult, and this mismatch is strongly dimension-dependent. Consistency shows a pronounced LLM-hard shift, whereas coherence shows a human-hard shift. We further show that human-easy but LLM-hard cases are partially predictable from observable source--summary properties. These findings suggest that aggregate human alignment reflects only part of LLM-as-a-judge reliability, while psychometric residual diagnostics support more informative judge evaluation and more targeted human--LLM collaboration.