ClaimDB : A Fact Verification Benchmark over Large Structured Data
Michael Theologitis, Preetam Prabhu Srikar Dammu, Chirag Shah, Dan Suciu · 2026
Read ClaimDB : A Fact Verification Benchmark over Large Structured Data by Michael Theologitis, Preetam Prabhu Srikar Dammu, Chirag Shah, Dan Suciu (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-2601-14698.txt.
Description
Michael Theologitis, Preetam Prabhu Srikar Dammu, Chirag Shah, and Dan Suciu build a fact-checking benchmark in this 2026 paper. Fact-verification benchmarks have made substantial progress. Claims grounded in large-scale structured data remain underexplored, though. Most benchmarks check claims against text passages. Real data claims rest on tables.
Their benchmark, ClaimDB, is the first fact-verification benchmark where evidence derives from compositions of millions of records and multiple tables. A claim may need sums, joins, and filters across a whole database to settle. Readers get a test that grades systems on database-grounded truth, a much harder bar than passage matching.
Catalog record
| CREATOR | Michael Theologitis, Preetam Prabhu Srikar Dammu, Chirag Shah, Dan Suciu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 53 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-14698