Logical Reasoning on LogicalDeduction (LD) (test)
85AccuracyDetermLR
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DetermLRModel=GPT-42023.10 | 85 | 12.35 | |
| CRModel=GPT-42023.10 | 78.33 | 16.98 | |
| LAMBADAModel=GPT-42023.10 | 78 | 67.32 | |
| ToTModel=GPT-42023.10 | 76.83 | 21.83 | |
| SIModel=GPT-42023.10 | 76.33 | 14.95 | |
| COT-SCModel=GPT-4, n=162023.10 | 74.67 | 16 | |
| DetermLRModel=GPT-3.5-turbo2023.10 | 74.33 | 13.19 | |
| COTModel=GPT-42023.10 | 73.33 | 1 | |
| StandardModel=GPT-42023.10 | 71.33 | 1 | |
| CRModel=GPT-3.5-turbo2023.10 | 71 | 18.32 | |
| ToTModel=GPT-3.5-turbo2023.10 | 66.33 | 23.71 | |
| LAMBADAModel=GPT-3.5-turbo2023.10 | 62.67 | 74.43 | |
| SIModel=GPT-3.5-turbo2023.10 | 51 | 17.24 | |
| COT-SCModel=GPT-3.5-turbo, n=162023.10 | 50.33 | 16 | |
| COTModel=GPT-3.5-turbo2023.10 | 46 | 1 | |
| StandardModel=GPT-3.5-turbo2023.10 | 41.33 | 1 |