Question Answering on ARC Easy (Accuracy, Std, Speedup)
93.2AccuracyTALE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TALEBackbone=Qwen 2.5 7B, Evaluation Protocol=0-shot, Layers Removed (#D)=32025.10 | 93.2 | 16 | -10 | |
| TALEBackbone=LLaMA 3.1 8B, Evaluation Protocol=0-shot, Layers Removed (#D)=52025.10 | 90.4 | 14 | -14.6 | |
| BSBABackbone=Qwen 2.5 7B, Evaluation Protocol=0-shot, Layers Removed (#D)=72025.10 | 90.08 | — | -30.3 | |
| BaselineBackbone=Qwen 2.5 7B, Evaluation Protocol=0-shot, Layers Removed (#D)=02025.10 | 90.04 | — | — | |
| BaselineBackbone=LLaMA 3.1 8B, Evaluation Protocol=0-shot, Layers Removed (#D)=02025.10 | 89 | — | — | |
| BSBABackbone=LLaMA 3.1 8B, Evaluation Protocol=0-shot, Layers Removed (#D)=82025.10 | 88.8 | — | -23.5 | |
| TALEBackbone=Mistral 7B, Evaluation Protocol=0-shot, Layers Removed (#D)=52025.10 | 85.6 | 14 | -15.4 | |
| BaselineBackbone=Mistral 7B, Evaluation Protocol=0-shot, Layers Removed (#D)=02025.10 | 83.8 | — | — | |
| BSBABackbone=Mistral 7B, Evaluation Protocol=0-shot, Layers Removed (#D)=92025.10 | 82.8 | — | -27.7 | |
| TALEBackbone=Lucie 7B, Evaluation Protocol=0-shot, Layers Removed (#D)=62025.10 | 75.8 | 15 | -18.1 | |
| BaselineBackbone=Lucie 7B, Evaluation Protocol=0-shot, Layers Removed (#D)=02025.10 | 74.4 | — | — | |
| BSBABackbone=Lucie 7B, Evaluation Protocol=0-shot, Layers Removed (#D)=82025.10 | 73.8 | — | -23.5 |