Math Reasoning on AIME 2024 (mean@16)
42Mean@16IRDS
Evaluation Results
| Method | Links | |
|---|---|---|
| IRDSModel=QWEN3-4B, Data Budget=20%2026.05 | 42 | |
| LIMRModel=QWEN3-4B, Data Budget=20%2026.05 | 30.6 | |
| PPL-TOPModel=QWEN3-4B, Data Budget=20%2026.05 | 28.6 | |
| DEPOModel=QWEN3-4B, Data Budget=20%2026.05 | 28.5 | |
| IFDModel=QWEN3-4B, Data Budget=20%2026.05 | 28.4 | |
| PPL-MIDDLEModel=QWEN3-4B, Data Budget=20%2026.05 | 27.9 | |
| IRDSModel=QWEN3-1.7B, Data Budget=20%2026.05 | 23.3 | |
| RANDOMModel=QWEN3-4B, Data Budget=20%2026.05 | 22.7 | |
| TOKEN-LENGTHModel=QWEN3-4B, Data Budget=20%2026.05 | 18.4 | |
| LIMRModel=QWEN3-1.7B, Data Budget=20%2026.05 | 17.2 | |
| DEPOModel=QWEN3-1.7B, Data Budget=20%2026.05 | 14.5 | |
| RANDOMModel=QWEN3-1.7B, Data Budget=20%2026.05 | 13.9 | |
| PPL-MIDDLEModel=QWEN3-1.7B, Data Budget=20%2026.05 | 12.3 | |
| PPL-TOPModel=QWEN3-1.7B, Data Budget=20%2026.05 | 12.2 | |
| IFDModel=QWEN3-1.7B, Data Budget=20%2026.05 | 11.7 | |
| TOKEN-LENGTHModel=QWEN3-1.7B, Data Budget=20%2026.05 | 10.7 | |
| DEPOModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 4.1 | |
| IRDSModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 3.4 | |
| IFDModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 3.2 | |
| TOKEN-LENGTHModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 3.1 | |
| PPL-MIDDLEModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 3.1 | |
| RANDOMModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 3 | |
| PPL-TOPModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 2.7 | |
| LIMRModel=LLAMA-3.1-8B-INSTRUCT, Data Budget=20%2026.05 | 2.4 |