SciTaRC: Benchmarking QA on Scientific Tabular Data that Requires Language Reasoning and Complex Computation
Hexuan Wang Yaxuan Ren Srikar Bommireddypalli Shuxian Chen · 2026
Read SciTaRC: Benchmarking QA on Scientific Tabular Data that Requires Language Reasoning and Complex Computation by Hexuan Wang Yaxuan Ren Srikar Bommireddypalli Shuxian Chen (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-2603-08910.txt.
Description
Hexuan Wang and colleagues build SciTaRC, a benchmark for question answering over scientific tables. Each question ships with a hand-built plan and complexity scores, enabling diagnosis beyond final accuracy. Top models fail on at least 23 percent of questions, and even oracle plans give only limited gains, exposing an execution bottleneck in long computational chains.
Catalog record
| CREATOR | Hexuan Wang Yaxuan Ren Srikar Bommireddypalli Shuxian Chen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37358 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.08910