Math Reasoning on GSM8K, MATH, SVAMP, ASDiv, MAWPS, TAB, MQA, MMLU STEM, and SAT (test)
40.4GSM8K AccuracyBLADE
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| BLADEBackbone=Llama2-7B, Train Toks.=3B, Few-shot CoT=true2026.06 | 40.4 | 9.8 | 60.8 | 66.1 | 84.9 | 49.6 | 27.3 | 33.1 | 21.9 | 43.8 | 2,375.1 | |
| RHO-1Backbone=Llama2-7B, Train Toks.=3B, Implementation=re-implemented, Few-shot CoT=true2026.06 | 33 | 7.6 | 57.9 | 60.6 | 78.9 | 48.4 | 24.9 | 30.1 | 28.1 | 41.1 | 2,140.7 | |
| BLADEBackbone=TinyLlama-1.1B, Train Toks.=3B, Few-shot CoT=true2026.06 | 20.8 | 8.6 | 35.7 | 48.9 | 66.5 | 21.9 | 17.2 | 22.4 | 25 | 29.7 | 494.3 | |
| Base-CTBackbone=Llama2-7B, Train Toks.=5B, Few-shot CoT=true2026.06 | 18.1 | 4.4 | 36.4 | 52.7 | 66.1 | 37.3 | 18.5 | 33.4 | 37.5 | 33.8 | 1,670.3 | |
| RandomBackbone=Llama2-7B, Train Toks.=3B, Few-shot CoT=true2026.06 | 17.3 | 5.2 | 36.9 | 51.9 | 64.2 | 40.2 | 17.1 | 32.9 | 25 | 32.3 | 1,011 | |
| MATESBackbone=Llama2-7B, Train Toks.=3B, Sample-level selection=true, Few-shot CoT=true2026.06 | 16.5 | 4.4 | 35.5 | 51.6 | 65.7 | 38.6 | 15.8 | 33.6 | 38.6 | 33.4 | 3,274.7 | |
| RHO-1Backbone=TinyLlama-1.1B, Train Toks.=3B, Implementation=re-implemented, Few-shot CoT=true2026.06 | 14.6 | 4.8 | 34.9 | 46.7 | 61.1 | 17 | 16 | 20.8 | 15.6 | 25.7 | 431.5 | |
| BaseBackbone=Llama2-7B, Few-shot CoT=true2026.06 | 14 | 3.6 | 39.5 | 51.7 | 63.5 | 30.9 | 12.4 | 32.7 | 34.4 | 31.4 | — | |
| MATESBackbone=TinyLlama-1.1B, Train Toks.=3B, Sample-level selection=true, Few-shot CoT=true2026.06 | 4.2 | 4.8 | 17.4 | 28.1 | 38.3 | 16 | 11.9 | 13.6 | 25 | 17.7 | 831.5 | |
| Base-CTBackbone=TinyLlama-1.1B, Train Toks.=5B, Few-shot CoT=true2026.06 | 3.7 | 3.6 | 17.6 | 27.8 | 35 | 17.1 | 9.1 | 18 | 18.8 | 16.7 | 314.6 | |
| RandomBackbone=TinyLlama-1.1B, Train Toks.=3B, Few-shot CoT=true2026.06 | 3.6 | 3.6 | 14.6 | 23.7 | 30.1 | 12.1 | 11.4 | 18.8 | 14.7 | 14.7 | 188.8 | |
| BaseBackbone=TinyLlama-1.1B, Few-shot CoT=true2026.06 | 2.9 | 3.2 | 11 | 18.1 | 20.4 | 12.5 | 14.6 | 16.1 | 21.9 | 13.4 | — |