Saved in:
Bibliographic Details
Main Authors: Huang, Tianyi, Huang, Nathan, Tang, Justin, Chen, Wenqian, Fan, Elsa
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.20562
Tags: Add Tag
No Tags, Be the first to tag this record!
Table of Contents:
  • Large language models (LLMs) are now widely used as judges, yet their decisions can change under presentation choices that should be irrelevant. We study one such source of instability: candidate-order sensitivity in listwise factuality evaluation, where several answers can look similarly polished while differing substantially in hallucination risk. We introduce PCFJudge, an inference-time method that reruns the same factuality-first listwise prompt over multiple orderings of the same candidate set and aggregates the resulting scores, ranks, and uncertainty signals into a single consensus decision. On RewardBench 2 Factuality, the final seven-permutation aggregate (K=7) improves top-1 selection accuracy from 86.00% to 91.33% with GPT-5.4 and from 86.33% to 89.67% with Claude Sonnet 4.6. These results suggest that candidate order can be a meaningful source of factuality-judging error and that marginalizing over this nuisance variation can improve the reliability of LLM evaluation.