Mathematical Reasoning on AIME25 (mean@16)
37.4Mean@16 ScoreIRDS
Evaluation Results
| Method | Links | |
|---|---|---|
| IRDSModel=QWEN3-4B, Data Budget=30%2026.05 | 37.4 | |
| DEPOModel=QWEN3-4B, Data Budget=30%2026.05 | 29.8 | |
| IFDModel=QWEN3-4B, Data Budget=30%2026.05 | 29.6 | |
| LIMRModel=QWEN3-4B, Data Budget=30%2026.05 | 29.4 | |
| PPL-TOPModel=QWEN3-4B, Data Budget=30%2026.05 | 29 | |
| PPL-MIDDLEModel=QWEN3-4B, Data Budget=30%2026.05 | 26.9 | |
| RANDOMModel=QWEN3-4B, Data Budget=30%2026.05 | 23.6 | |
| IRDSModel=QWEN3-1.7B, Data Budget=30%2026.05 | 23.4 | |
| TOKEN-LENGTHModel=QWEN3-4B, Data Budget=30%2026.05 | 20.4 | |
| DEPOModel=QWEN3-1.7B, Data Budget=30%2026.05 | 16.1 | |
| RANDOMModel=QWEN3-1.7B, Data Budget=30%2026.05 | 15.2 | |
| LIMRModel=QWEN3-1.7B, Data Budget=30%2026.05 | 14.1 | |
| PPL-MIDDLEModel=QWEN3-1.7B, Data Budget=30%2026.05 | 13.2 | |
| TOKEN-LENGTHModel=QWEN3-1.7B, Data Budget=30%2026.05 | 12.6 | |
| IFDModel=QWEN3-1.7B, Data Budget=30%2026.05 | 11.4 | |
| PPL-TOPModel=QWEN3-1.7B, Data Budget=30%2026.05 | 10.2 | |
| RANDOMModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=30%2026.05 | 2 | |
| IFDModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=30%2026.05 | 1.8 | |
| DEPOModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=30%2026.05 | 1.6 | |
| PPL-MIDDLEModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=30%2026.05 | 0.9 | |
| LIMRModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=30%2026.05 | 0.9 | |
| IRDSModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=30%2026.05 | 0.9 | |
| TOKEN-LENGTHModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=30%2026.05 | 0.7 | |
| PPL-TOPModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=30%2026.05 | 0.3 |