Reasoning on MGSM
94.4AccuracyeMoT
Evaluation results
| Method | Links | |
|---|---|---|
| eMoTModels=Qwen-32B2026.06 | 94.4 | |
| RoTBackbone=GPT-4, Shots=2-shot2024.10 | 90 | |
| BoTModels=GPT-42026.06 | 89.3 | |
| RoTBackbone=GPT-4, Shots=1-shot2024.10 | 88.7 | |
| BOTBackbone=GPT-4, Shots=1-shot2024.10 | 87.9 | |
| BOTBackbone=GPT-4, Shots=2-shot2024.10 | 87.5 | |
| GOTBackbone=GPT-4, Shots=2-shot2024.10 | 86.7 | |
| TOTBackbone=GPT-4, Shots=2-shot2024.10 | 86.3 | |
| GOTBackbone=GPT-4, Shots=1-shot2024.10 | 85.9 | |
| Meta-promptingBackbone=GPT-4, Shots=2-shot2024.10 | 85.5 | |
| Meta-promptingBackbone=GPT-4, Shots=1-shot2024.10 | 85.1 | |
| RoTBackbone=GPT-3.5-turbo, Shots=2-shot2024.10 | 84.7 | |
| TOTBackbone=GPT-4, Shots=1-shot2024.10 | 84.7 | |
| COTBackbone=GPT-4, Shots=2-shot2024.10 | 84.3 | |
| COTBackbone=GPT-4, Shots=1-shot2024.10 | 83.9 | |
| BOTBackbone=GPT-3.5-turbo, Shots=2-shot2024.10 | 82.3 | |
| RoTBackbone=GPT-3.5-turbo, Shots=1-shot2024.10 | 82.3 | |
| Meta-promptingBackbone=GPT-3.5-turbo, Shots=2-shot2024.10 | 81.5 | |
| BOTBackbone=GPT-3.5-turbo, Shots=1-shot2024.10 | 81.1 | |
| GOTBackbone=GPT-3.5-turbo, Shots=2-shot2024.10 | 81 | |
| Meta-promptingBackbone=GPT-3.5-turbo, Shots=1-shot2024.10 | 80.3 | |
| GOTBackbone=GPT-3.5-turbo, Shots=1-shot2024.10 | 80.3 | |
| TOTBackbone=GPT-3.5-turbo, Shots=2-shot2024.10 | 79.4 | |
| TOTBackbone=GPT-3.5-turbo, Shots=1-shot2024.10 | 78.3 | |
| COTBackbone=GPT-3.5-turbo, Shots=2-shot2024.10 | 77.4 | |
| COTBackbone=GPT-3.5-turbo, Shots=1-shot2024.10 | 75 | |
| PaLModels=Codex (175B)2026.06 | 73.6 | |
| Qwen-32B (Direct)Models=Qwen-32B2026.06 | 48 | |
| ToTModels=GPT-42026.06 | 32.8 |