Table Reasoning on BIRD (test)
29AccuracyCPO
Evaluation Results
| Method | Links | |
|---|---|---|
| CPOModel=Qwen3-14B-Base2026.06 | 29 | |
| CPOModel=Mistral-8B-Instruct2026.06 | 26 | |
| Trace SFTModel=Qwen3-14B-Base2026.06 | 23 | |
| CPOModel=Llama-3.1-8B-Instruct2026.06 | 23 | |
| Q&A SFTModel=Qwen3-14B-Base2026.06 | 22 | |
| Trace SFTModel=Llama-3.1-8B-Instruct2026.06 | 18 | |
| Q&A SFTModel=Mistral-8B-Instruct2026.06 | 17 | |
| Q&A SFTModel=Llama-3.1-8B-Instruct2026.06 | 16 | |
| Trace SFTModel=Mistral-8B-Instruct2026.06 | 12 | |
| Table-R1Model=Table-R12026.06 | 10.3 | |
| DPOModel=Mistral-8B-Instruct2026.06 | 9 | |
| DPOModel=Qwen3-14B-Base2026.06 | 7.8 | |
| DPOModel=Llama-3.1-8B-Instruct2026.06 | 6.9 |