Science Question Answering on ARC Challenge (Accuracy and Exit Position)
62.3AccuracyMinistral3 8B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Ministral3 8BModel Architecture=Ministral3 8B (34 layers), Evaluation Protocol=Backbone2026.04 | 62.3 | — | |
| River-LLMModel Architecture=Ministral3 8B (34 layers), Early-exit threshold (tau)=0.8, Evaluation Protocol=Adaptive2026.04 | 62.2 | 2.01 | |
| River-LLMModel Architecture=Ministral3 8B (34 layers), Early-exit threshold (tau)=0.9, Evaluation Protocol=Adaptive2026.04 | 62.2 | 2.11 | |
| Phi4-miniModel Architecture=Phi4-mini (32 layers), Evaluation Protocol=Backbone2026.04 | 57.1 | — | |
| River-LLMModel Architecture=Phi4-mini (32 layers), Early-exit threshold (tau)=0.9, Evaluation Protocol=Adaptive2026.04 | 54.4 | 3.45 | |
| River-LLMModel Architecture=Phi4-mini (32 layers), Early-exit threshold (tau)=0.8, Evaluation Protocol=Adaptive2026.04 | 54 | 2.02 |