Assessing Y-Axis Influence: Bias in Multimodal Language Models on Chart-to-Table Translation
Seok Hwan Song, Azher Ahmed Efat, Wallapak Tavanapong · 2026
Read Assessing Y-Axis Influence: Bias in Multimodal Language Models on Chart-to-Table Translation by Seok Hwan Song, Azher Ahmed Efat, Wallapak Tavanapong (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-2604-24987.txt.
Description
Song, Efat and Tavanapong examine y-axis bias in multimodal language models on chart-to-table translation. Chart-to-table translation turns chart images into tabular data. They find imbalances in y-axis information across public chart datasets, which can bias models and cause uneven performance. Their FairChart2Table framework analyses y-axis bias on five leading models. Biases tie to tick-value digit length, tick count, value range and tick format. Legend and entity counts also affect performance, while prompting with y-axis information helps some models.
Catalog record
| CREATOR | Seok Hwan Song, Azher Ahmed Efat, Wallapak Tavanapong |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 35549 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.24987