Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2601.13600 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909995544608768 |
|---|---|
| author | He, Paul Kirschbaum, Elke Kasiviswanathan, Shiva |
| author_facet | He, Paul Kirschbaum, Elke Kasiviswanathan, Shiva |
| contents | Ensuring that collections of natural-language facts are globally consistent is essential for tasks such as fact-checking, summarization, and knowledge base construction. While Large Language Models (LLMs) can assess the consistency of small subsets of facts, their judgments are noisy, and pairwise checks are insufficient to guarantee global coherence. We formalize this problem and show that verifying global consistency requires exponentially many oracle queries in the worst case. To make the task practical, we propose an adaptive divide-and-conquer algorithm that identifies minimal inconsistent subsets (MUSes) of facts and optionally computes minimal repairs through hitting-sets. Our approach has low-degree polynomial query complexity. Experiments with both synthetic and real LLM oracles show that our method efficiently detects and localizes inconsistencies, offering a scalable framework for linguistic consistency verification with LLM-based evaluators. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_13600 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Foundations of Global Consistency Checking with Noisy LLM Oracles He, Paul Kirschbaum, Elke Kasiviswanathan, Shiva Artificial Intelligence Ensuring that collections of natural-language facts are globally consistent is essential for tasks such as fact-checking, summarization, and knowledge base construction. While Large Language Models (LLMs) can assess the consistency of small subsets of facts, their judgments are noisy, and pairwise checks are insufficient to guarantee global coherence. We formalize this problem and show that verifying global consistency requires exponentially many oracle queries in the worst case. To make the task practical, we propose an adaptive divide-and-conquer algorithm that identifies minimal inconsistent subsets (MUSes) of facts and optionally computes minimal repairs through hitting-sets. Our approach has low-degree polynomial query complexity. Experiments with both synthetic and real LLM oracles show that our method efficiently detects and localizes inconsistencies, offering a scalable framework for linguistic consistency verification with LLM-based evaluators. |
| title | Foundations of Global Consistency Checking with Noisy LLM Oracles |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2601.13600 |