Table Question Answering on TAT-QA (Execution Match)
77.78Execution Match (EM)GPT-4o
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-4oModel Type=Base Model2025.03 | 77.78 | |
| Qwen2.5 14BModel Type=Base Model2025.03 | 15.47 | |
| Qwen2.5 14BModel Type=FT_Syn-QA2025.03 | 15.03 | |
| Qwen2.5-coder 14BModel Type=FT_Syn-QA2025.03 | 14.6 | |
| Qwen2.5-coder 14BModel Type=Base Model2025.03 | 14.37 | |
| Qwen2.5 3BModel Type=FT_Syn-QA2025.03 | 11.11 | |
| Qwen2.5-coder 3BModel Type=FT_Syn-QA2025.03 | 8.06 | |
| CodeLlama-Instruct 13BModel Type=FT_Syn-QA2025.03 | 7.41 | |
| Qwen2.5 3BModel Type=Base Model2025.03 | 6.75 | |
| Llama2 13BModel Type=FT_Syn-QA2025.03 | 6.54 | |
| Qwen2.5-coder 3BModel Type=Base Model2025.03 | 5.88 | |
| Llama2 7BModel Type=FT_Syn-QA2025.03 | 3.92 | |
| CodeLlama-Instruct 13BModel Type=Base Model2025.03 | 3.49 | |
| CodeLlama-Instruct 7BModel Type=FT_Syn-QA2025.03 | 2.18 | |
| Llama2 13BModel Type=Base Model2025.03 | 1.09 | |
| CodeLlama-Instruct 7BModel Type=Base Model2025.03 | 0.44 | |
| Llama2 7BModel Type=Base Model2025.03 | 0 |