Mathematical Reasoning on GSM8K (Score, #Strategy)
92.9ScorePrompting
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PromptingReasoner=Llama3.3-70B, Planner=70B2025.12 | 92.9 | 4 | |
| rSIM (Plug-in)Reasoner=Llama3.3-70B, Planner=3B plug-in2025.12 | 92.1 | 2 | |
| rSIM (Transfer)Reasoner=Llama3.3-70B, Planner=Qwen2.5-1.5B2025.12 | 92 | 2 | |
| rSIM (Plug-in)Reasoner=Llama3.3-70B, Planner=1B plug-in2025.12 | 91.7 | 1 | |
| PS+Reasoner=Llama3.3-70B, Planner=No2025.12 | 90.9 | 0 | |
| ZeroCoTReasoner=Llama3.3-70B, Planner=No2025.12 | 90.5 | 0 | |
| PromptingReasoner=Llama3.3-70B, Planner=3B2025.12 | 90.5 | 4 | |
| rSIMReasoner=Llama3.2-1B, Planner=3B2025.12 | 86.3 | 3 | |
| rSIMReasoner=Llama3.2-1B, Planner=Qwen2.5-1.5B2025.12 | 84.4 | 1 | |
| rSIMReasoner=Llama3.2-1B, Planner=1B2025.12 | 83.9 | 1 | |
| PromptingReasoner=Llama3.2-3B, Planner=70B2025.12 | 81.8 | 4 | |
| ZeroCoTReasoner=Llama3.2-3B, Planner=No2025.12 | 77.7 | 0 | |
| PS+Reasoner=Llama3.2-3B, Planner=No2025.12 | 77.7 | 0 | |
| PromptingReasoner=Llama3.2-3B, Planner=3B2025.12 | 77.1 | 5 | |
| PromptingReasoner=Llama3.2-1B, Planner=70B2025.12 | 46.9 | 3 | |
| ZeroCoTReasoner=Llama3.2-1B, Planner=No2025.12 | 44.4 | 0 | |
| PS+Reasoner=Llama3.2-1B, Planner=No2025.12 | 43.7 | 0 | |
| PromptingReasoner=Llama3.2-1B, Planner=3B2025.12 | 42.6 | 3 |