Overall Table Generation on ARXIV2TABLE
56.7Average Scoreiterative batch-based generation method
Evaluation Results
| Method | Links | |
|---|---|---|
| iterative batch-based generation methodBackbone Model=GPT-o32025.04 | 56.7 | |
| iterative batch-based generation methodBackbone Model=GPT-4o, Chain of Thought=true2025.04 | 55 | |
| iterative batch-based generation methodBackbone Model=GPT-4o2025.04 | 53.3 | |
| iterative batch-based generation methodBackbone Model=GPT-4o-mini2025.04 | 52 | |
| iterative batch-based generation methodBackbone Model=Mistral-Large (123B)2025.04 | 48.8 | |
| iterative batch-based generation methodBackbone Model=DeepSeek-V3 (685B)2025.04 | 48.8 | |
| iterative batch-based generation methodBackbone Model=LLaMa-3.3 (70B)2025.04 | 46.6 | |
| Newman et al.Backbone Model=GPT-4o, Chain of Thought=true2025.04 | 44.5 | |
| Newman et al.Backbone Model=GPT-4o2025.04 | 42.8 | |
| Newman et al.Backbone Model=GPT-4o-mini2025.04 | 40.5 | |
| Newman et al.Backbone Model=DeepSeek-V3 (685B)2025.04 | 39.5 | |
| Newman et al.Backbone Model=Mistral-Large (123B)2025.04 | 37.6 | |
| Baseline 2Backbone Model=DeepSeek-V3 (685B)2025.04 | 37.2 | |
| Baseline 1Backbone Model=GPT-4o2025.04 | 36.7 | |
| Baseline 1Backbone Model=DeepSeek-V3 (685B)2025.04 | 35.8 | |
| Baseline 2Backbone Model=GPT-4o2025.04 | 35.8 | |
| Newman et al.Backbone Model=LLaMa-3.3 (70B)2025.04 | 35.3 | |
| Baseline 2Backbone Model=GPT-4o-mini2025.04 | 34.5 | |
| Baseline 1Backbone Model=LLaMa-3.3 (70B)2025.04 | 32.8 | |
| Baseline 1Backbone Model=GPT-4o-mini2025.04 | 31.6 | |
| Baseline 2Backbone Model=Mistral-Large (123B)2025.04 | 30.6 | |
| Baseline 1Backbone Model=Mistral-Large (123B)2025.04 | 27.9 | |
| Baseline 2Backbone Model=LLaMa-3.3 (70B)2025.04 | 26.6 |