Math Reasoning on Minerva (mean@16)
43.8Mean@16IRDS
Evaluation Results
| Method | Links | |
|---|---|---|
| IRDSModel=QWEN3-4B, Data Budget=20%2026.05 | 43.8 | |
| PPL-MIDDLEModel=QWEN3-4B, Data Budget=20%2026.05 | 42.9 | |
| LIMRModel=QWEN3-4B, Data Budget=20%2026.05 | 42.7 | |
| DEPOModel=QWEN3-4B, Data Budget=20%2026.05 | 41.8 | |
| PPL-TOPModel=QWEN3-4B, Data Budget=20%2026.05 | 40.8 | |
| IFDModel=QWEN3-4B, Data Budget=20%2026.05 | 40.3 | |
| RANDOMModel=QWEN3-4B, Data Budget=20%2026.05 | 39.4 | |
| TOKEN-LENGTHModel=QWEN3-4B, Data Budget=20%2026.05 | 38.9 | |
| IRDSModel=QWEN3-1.7B, Data Budget=20%2026.05 | 33.6 | |
| LIMRModel=QWEN3-1.7B, Data Budget=20%2026.05 | 31.7 | |
| DEPOModel=QWEN3-1.7B, Data Budget=20%2026.05 | 31.1 | |
| RANDOMModel=QWEN3-1.7B, Data Budget=20%2026.05 | 30 | |
| PPL-TOPModel=QWEN3-1.7B, Data Budget=20%2026.05 | 29.8 | |
| PPL-MIDDLEModel=QWEN3-1.7B, Data Budget=20%2026.05 | 29.5 | |
| IFDModel=QWEN3-1.7B, Data Budget=20%2026.05 | 28.3 | |
| IRDSModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 28.3 | |
| DEPOModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 27.4 | |
| TOKEN-LENGTHModel=QWEN3-1.7B, Data Budget=20%2026.05 | 27.2 | |
| PPL-MIDDLEModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 25.9 | |
| LIMRModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 25.7 | |
| RANDOMModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 25.3 | |
| TOKEN-LENGTHModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 25.2 | |
| IFDModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 25 | |
| PPL-TOPModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 24.3 |