Chain-of-Thought Generation on SVAMP (test)
4.63GPT-4 ScoreGold
Evaluation Results
| Method | Links | |
|---|---|---|
| GoldSelf-consistency=false2024.03 | 4.63 | |
| GoldSelf-consistency=true2024.03 | 4.43 | |
| MI-based distillationSelf-consistency=true2024.03 | 2.72 | |
| DSSSelf-consistency=true2024.03 | 2.53 | |
| DSSSelf-consistency=false2024.03 | 2.5 | |
| MI-based distillationSelf-consistency=false2024.03 | 2.3 |