Mathematical Reasoning on GSM8K 500 samples (test)
2,578.42Avg TokensPostCal
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PostCalLLM Backbone=DeepSeek-R1-Distill-Qwen-1.5B2026.06 | 2,578.42 | 87.07 | |
| ITCRLLM Backbone=DeepSeek-R1-Distill-Qwen-1.5B2026.06 | 2,559.58 | 83.94 | |
| PostCalLLM Backbone=LLaMA-3.1-8B-Instruct2026.06 | 2,092.8 | 67.69 | |
| PostCalLLM Backbone=Qwen3-4B-Thinking-25072026.06 | 1,975.56 | 64.82 | |
| ITCRLLM Backbone=LLaMA-3.1-8B-Instruct2026.06 | 1,919.5 | 62.4 | |
| ITCRLLM Backbone=Qwen3-4B-Thinking-25072026.06 | 1,812.92 | 55.04 |