List your company

Science Question Answering on ARC-C (Accuracy)

51.47Accuracy (ARC-C)Pre-trained
22.97430.37237.7745.168May 25, 2026May 26, 2026May 27, 2026May 28, 2026May 29, 2026May 30, 2026Jun 1, 2026

Evaluation results

MethodLinks
Pre-trainedModel=Gemma-3-4B2026.0651.47
AlphaTokenModel=Gemma-3-4B2026.0650.74
Token CleaningModel=Gemma-3-4B2026.0650.15
LESSModel=Gemma-3-4B2026.0649.97
ssTOKENModel=Gemma-3-4B2026.0649.74
STMModel=Gemma-3-4B2026.0649.53
LoRAModel=Gemma-3-4B2026.0648.96
XTFModel=Gemma-3-4B2026.0648.56
Standard FTModel=Gemma-3-4B2026.0643.52
HybKDDistillation Pair=Llama3.1-8B → Llama3.2-1B2026.0537.01
α-β divergenceDistillation Pair=Llama3.1-8B → Llama3.2-1B2026.0534.9
JS divergenceDistillation Pair=Llama3.1-8B → Llama3.2-1B2026.0534.42
Adaptive KLDistillation Pair=Llama3.1-8B → Llama3.2-1B2026.0533.74
Skew FKLDistillation Pair=Llama3.1-8B → Llama3.2-1B2026.0532.34
Skew RKLDistillation Pair=Llama3.1-8B → Llama3.2-1B2026.0532.24
Reverse KLDistillation Pair=Llama3.1-8B → Llama3.2-1B2026.0532.08
Total VariationDistillation Pair=Llama3.1-8B → Llama3.2-1B2026.0528.99
Reverse KLDistillation Pair=Gemma3-4B → Gemma3-1B2026.0527.58
Skew FKLDistillation Pair=Gemma3-4B → Gemma3-1B2026.0526.14
α-β divergenceDistillation Pair=Gemma3-4B → Gemma3-1B2026.0526.07
HybKDDistillation Pair=Gemma3-4B → Gemma3-1B2026.0526.07
JS divergenceDistillation Pair=Gemma3-4B → Gemma3-1B2026.0526.03
Skew RKLDistillation Pair=Gemma3-4B → Gemma3-1B2026.0525.93
Total VariationDistillation Pair=Gemma3-4B → Gemma3-1B2026.0525.83
Adaptive KLDistillation Pair=Gemma3-4B → Gemma3-1B2026.0524.07