Table Reasoning on WikiTQ (Exact Match (EM))
79.4Exact Match (EM)TabTrim-8B
Evaluation Results
| Method | Links | |
|---|---|---|
| TabTrim-8BMethod Category=Ours, Backbone/Model=TabTrim-8B, Pipeline Reasoner=GPT-4o-mini2026.01 | 79.4 | |
| TabTrim-4BMethod Category=Ours, Backbone/Model=TabTrim-4B, Pipeline Reasoner=GPT-4o-mini2026.01 | 76.8 | |
| Table-CriticMethod Category=Critique methods, Pipeline Reasoner=GPT-4o-mini2026.01 | 72.6 | |
| TALONMethod Category=Critique methods, Pipeline Reasoner=GPT-4o-mini2026.01 | 70.7 | |
| TabSQLifyMethod Category=Program-based methods, Pipeline Reasoner=GPT-4o-mini2026.01 | 68.7 | |
| Chain-of-TableMethod Category=LLM-based methods, Pipeline Reasoner=GPT-4o-mini2026.01 | 67.5 | |
| DaterMethod Category=LLM-based methods, Pipeline Reasoner=GPT-4o-mini2026.01 | 65.8 | |
| BinderMethod Category=Program-based methods, Pipeline Reasoner=GPT-4o-mini2026.01 | 54.8 | |
| GPT-4o-miniMethod Category=Direct QA, Backbone/Model=GPT-4o-mini2026.01 | 54.3 | |
| Qwen3-8BMethod Category=Direct QA, Backbone/Model=Qwen3-8B2026.01 | 52.2 | |
| Qwen3-4BMethod Category=Direct QA, Backbone/Model=Qwen3-4B2026.01 | 50.8 |