List your company

Science Question Answering on ARC Challenge (Accuracy and Exit Position)

62.3AccuracyMinistral3 8B
53.66855.90958.1560.391Apr 20, 2026

Evaluation results

MethodLinks
Ministral3 8BModel Architecture=Ministral3 8B (34 layers), Evaluation Protocol=Backbone2026.0462.3—
River-LLMModel Architecture=Ministral3 8B (34 layers), Early-exit threshold (tau)=0.8, Evaluation Protocol=Adaptive2026.0462.22.01
River-LLMModel Architecture=Ministral3 8B (34 layers), Early-exit threshold (tau)=0.9, Evaluation Protocol=Adaptive2026.0462.22.11
Phi4-miniModel Architecture=Phi4-mini (32 layers), Evaluation Protocol=Backbone2026.0457.1—
River-LLMModel Architecture=Phi4-mini (32 layers), Early-exit threshold (tau)=0.9, Evaluation Protocol=Adaptive2026.0454.43.45
River-LLMModel Architecture=Phi4-mini (32 layers), Early-exit threshold (tau)=0.8, Evaluation Protocol=Adaptive2026.04542.02