VeriLLMed : Interactive Visual Debugging of Medical Large Language Models with Knowledge Graphs
Yurui Xiang, Xingyi Mao, Rui Sheng, Zixin Chen, Zelin Zang, Yuyang Wu, Haipeng Zeng, Huamin Qu, Yushi Sun*, and Yanna Lin* · 2026
Read VeriLLMed : Interactive Visual Debugging of Medical Large Language Models with Knowledge Graphs by Yurui Xiang, Xingyi Mao, Rui Sheng, Zixin Chen, Zelin Zang, Yuyang Wu, Haipeng Zeng, Huamin Qu, Yushi Sun*, and Yanna Lin* (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-23356.txt.
Description
Xiang and nine colleagues present VeriLLMed, a visual system for debugging medical models with biomedical knowledge graphs. Developers often lack medical expertise to read errors clinically and face diverse failures across inputs and tasks. They struggle to spot recurring patterns with instance-centric manual review. VeriLLMed turns outputs into matched reasoning paths, builds knowledge-graph references, and names three error classes: relation, branch, and missing errors. Case studies and expert review show it helps developers spot implausible reasoning and generate useful insights.
Catalog record
| CREATOR | Yurui Xiang, Xingyi Mao, Rui Sheng, Zixin Chen, Zelin Zang, Yuyang Wu, Haipeng Zeng, Huamin Qu, Yushi Sun*, and Yanna Lin* |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 72952 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.23356