Role-playing on Alpaca-P 50P x 100Q scale (test)
81.97Alpaca-P ScorePersona-Pruner
Evaluation Results
| Method | Links | |
|---|---|---|
| Persona-PrunerBackbone=Llama-3.2-3B-Instruct, Sparsity=25%2026.06 | 81.97 | |
| Adapt-PrunerBackbone=Llama-3.2-3B-Instruct, Sparsity=25%2026.06 | 63.42 | |
| LLM-PrunerBackbone=Llama-3.2-3B-Instruct, Sparsity=25%2026.06 | 61.91 | |
| Depth PruningBackbone=Llama-3.2-3B-Instruct, Sparsity=25%2026.06 | 21.48 | |
| SliceGPTBackbone=Llama-3.2-3B-Instruct, Sparsity=25%2026.06 | 5.42 |