Table-R1: Region-based Reinforcement Learning for Table Understanding

Benchmark Model Rank Results
table-question-answering-on-tablebenchGPT-4o (TCoT)–Score: 51.96
table-question-answering-on-tablebenchQWQ-32B (TCoT)–Score: 43.48
table-question-answering-on-tablebenchTable-R1 (Qwen2-7B, TCoT)–Score: 41.85
table-question-answering-on-tablebenchTableGPT2-72B (TCoT)–Score: 50.06