Contradiction to Consensus: Dual - Perspective, Multi - Source Retrieval-Based Claim Verification with Source - Level Disagreement using LLM
Md Badsha Biswas, Özlem Uzuner · 2026
Read Contradiction to Consensus: Dual - Perspective, Multi - Source Retrieval-Based Claim Verification with Source - Level Disagreement using LLM by Md Badsha Biswas, Özlem Uzuner (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-2602-18693.txt.
Description
Fact-checking tools often trust one source and ignore the rest, which hides exactly the disagreements a reader most needs to see. Md Badsha Biswas, with one co-author, builds a 2026 claim-verification system that gathers evidence for a claim and for its negation, from Wikipedia, PubMed, and Google.
The retrieved sets are filtered, deduplicated, and pooled, then fed to large language models for a verdict, with confidence scores used to show where sources disagree. Tests on four benchmarks with five models show the pooled evidence improves checking and makes source-specific reasoning visible. The point is not just better scores, but a more honest picture of contested information.
Catalog record
| CREATOR | Md Badsha Biswas, Özlem Uzuner |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 29 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-18693