Table Question Answering on WikiTQ (BLEU, ROUGE-L)
63.19BLEUGPT-5-nano (RE-TAB RG)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GPT-5-nano (RE-TAB RG)Workflow=Tree-of-Table (Tools), ST Reward=Enabled, SR Reward=RG2026.01 | 63.19 | 70.63 | |
| GPT-4.1-nano (RE-TAB RG)Workflow=Chain-of-Table (Tools), ST Reward=Enabled, SR Reward=RG2026.01 | 59.44 | 63.83 | |
| QWEN3-8B (RE-TAB AS)Workflow=Chain-of-Table (Tools), ST Reward=Enabled, SR Reward=AS2026.01 | 30 | 33.5 | |
| QWEN3-8B (RE-TAB RG)Workflow=Chain-of-Table (Tools), ST Reward=Enabled, SR Reward=RG2026.01 | 27.77 | 35.83 | |
| QWEN3-8B (RE-TAB AC)Workflow=Chain-of-Table (Tools), ST Reward=Enabled, SR Reward=AC2026.01 | 21.98 | 26.33 |