Where Cross-Question Structure in LLM Errors Comes From: Shared Seeds, Shared Prompts, and Answer-Letter Preference

2026-10-26 · ArXiv

arXiv:2607.20464v2 Announce Type: replace

Abstract: Disagreement between language models is often read as missing knowledge. Questions that need the same missing knowledge should then be right or wrong together. We test for this structure with the eigenvalues of the questions' correlation matrix. Within one model, independently drawn answers show no detectable structure; it appears when all questions in one pass over the benchmark share a random seed or few-shot prompt. Across 24 models, the largest structure beyond overall skill is a preference among answer letters: when the options are rotated, it follows the letter, not the question, and it depends on one model, Llama-3.2-1B; without it only a smaller excess remains. Counting a question correct only if right in every rotation removes the part that acts on guesses. What remains is a marginal excess at the 5% line, tied post hoc to two near-duplicate model pairs and to Llama-3.2-1B.

原文链接：https://arxiv.org/abs/2607.20464

查看更多资讯请见 AI 行业资讯。