Table Question Answering on WikiTableQuestions OOD
83.34AccuracySkillGrad
Evaluation results
| Method | Links | |
|---|---|---|
| SkillGradBackbone=GPT-5.4, Initialization=Third-party, Training Type=Training-based skill improvement2026.05 | 83.34 | |
| SkillGradBackbone=GPT-5.4, Initialization=LLM-gen., Training Type=Training-based skill improvement2026.05 | 82.38 | |
| Trace2SkillBackbone=GPT-5.4, Initialization=Third-party, Training Type=Training-based skill improvement2026.05 | 81.91 | |
| EvoSkillBackbone=GPT-5.4, Initialization=Third-party, Training Type=Training-based skill improvement2026.05 | 80.95 | |
| Trace2SkillBackbone=GPT-5.4, Initialization=LLM-gen., Training Type=Training-based skill improvement2026.05 | 79.05 | |
| No SkillBackbone=GPT-5.4, Initialization=None, Training Type=Training-free2026.05 | 78.57 | |
| Base xlsx SkillBackbone=GPT-5.4, Initialization=Third-party, Training Type=Training-free2026.05 | 78.57 | |
| EvoSkillBackbone=GPT-5.4, Initialization=LLM-gen., Training Type=Training-based skill improvement2026.05 | 78.09 | |
| Base xlsx SkillBackbone=GPT-5.4, Initialization=LLM-gen., Training Type=Training-free2026.05 | 77.14 | |
| SkillGradBackbone=GPT-4.1, Initialization=LLM-gen., Training Type=Training-based skill improvement2026.05 | 73.65 | |
| Trace2SkillBackbone=GPT-4.1, Initialization=LLM-gen., Training Type=Training-based skill improvement2026.05 | 60 | |
| SkillGradBackbone=GPT-4.1, Initialization=Third-party, Training Type=Training-based skill improvement2026.05 | 53.81 | |
| EvoSkillBackbone=GPT-4.1, Initialization=LLM-gen., Training Type=Training-based skill improvement2026.05 | 53.33 | |
| No SkillBackbone=GPT-4.1, Initialization=None, Training Type=Training-free2026.05 | 52.86 | |
| Trace2SkillBackbone=GPT-4.1, Initialization=Third-party, Training Type=Training-based skill improvement2026.05 | 51.43 | |
| Base xlsx SkillBackbone=GPT-4.1, Initialization=LLM-gen., Training Type=Training-free2026.05 | 48.58 | |
| EvoSkillBackbone=GPT-4.1, Initialization=Third-party, Training Type=Training-based skill improvement2026.05 | 43.33 | |
| Base xlsx SkillBackbone=GPT-4.1, Initialization=Third-party, Training Type=Training-free2026.05 | 42.86 |