TabScope: Question-Adaptive Scope Selection for Table Question Answering
Large Language Models (LLMs) have shown strong performance on table question answering, yet their accuracy often degrades as table size increases. We find that this degradation is not uniform across question types. Localization-sensitive questions are particularly affected by irrelevant table conten...
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Artículo |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2609.03395 http://arxiv.org/abs/2609.03395 |
| Etiquetas: |
Sin Etiquetas, Sea el primero en etiquetar este registro!
|