Fact Verification on SciTab
71Macro F1MACE
Evaluation Results
| Method | Links | |
|---|---|---|
| MACEBackbone=Qw-235B, Components=Pm + Em + Vm'2026.04 | 71 | |
| MACEBackbone=Qw-72B, Components=Pm + Em + Vm'2026.04 | 67 | |
| GPT-42026.04 | 65 | |
| MACEBackbone=Ll-8B, Components=Pm + Em + Vm'2026.04 | 65 | |
| TARTBackbone=GPT-42026.04 | 64 | |
| GPT-4 + COTprompting=Chain-of-Thought2026.04 | 63 | |
| MACEBackbone=Mt-7B, Components=Pm + Em + Vm'2026.04 | 58 | |
| ProTrix2026.04 | 43 | |
| InstructGPT + COTprompting=Chain-of-Thought2026.04 | 43 | |
| InstructGPT2026.04 | 42 | |
| Vicuna (13B)2026.04 | 35 | |
| PASTA2026.04 | 33 | |
| Ll-13B2026.04 | 33 | |
| Alpaca-7B2026.04 | 29 |