Abductive Reasoning on FOLIO
88Accuracy (FOLIO)PACS
Evaluation Results
| Method | Links | |
|---|---|---|
| PACSBackbone=Llama 3.3 70B2026.05 | 88 | |
| PACSBackbone=Llama 3-Instruct 8B2026.05 | 82 | |
| ARGOSBackbone=Llama 3.3 70B2026.05 | 80 | |
| ARGOSBackbone=Llama 3-Instruct 8B2026.05 | 78 | |
| If-BeamBackbone=Llama 3.3 70B2026.05 | 78 | |
| SC-20Backbone=Llama 3.3 70B2026.05 | 77 | |
| If-BeamBackbone=Llama 3-Instruct 8B2026.05 | 74 | |
| COTBackbone=Llama 3.3 70B2026.05 | 73 | |
| SC-20Backbone=Llama 3-Instruct 8B2026.05 | 71 | |
| LoTBackbone=Llama 3.3 70B2026.05 | 70 | |
| LoTBackbone=Llama 3-Instruct 8B2026.05 | 69 | |
| COTBackbone=Llama 3-Instruct 8B2026.05 | 68 | |
| LLM-TreeBackbone=Llama 3-Instruct 8B2026.05 | 63 | |
| LLM-TreeBackbone=Llama 3.3 70B2026.05 | 63 |