Zero-shot Reasoning on Multiple Reasoning Benchmarks Average
53.5Average AccuracyROCKET-default
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ROCKET-defaultModel=Llama-3.2-1B, Compression Ratio=20%2026.06 | 53.5 | — | |
| ROCKET-ActCostModel=Llama-3.2-1B, Compression Ratio=20%2026.06 | 53.3 | 0.2 |