akhauriyash/DeepSeek-R1-Distill-Llama-8B-Butler
Feature Extraction • 8B • Updated
• 1
TokenButler
TokenButler -- Predict token importance for all heads across the transformer in the first layer itself. Enable fine-grained token sparsity!
-
-
akhauriyash/Llama-3.1-8B-Butler
Text Generation • 8B • Updated • 1 -
akhauriyash/Llama-2-7b-hf-Butler
Text Generation • 7B • Updated • 2 -
akhauriyash/Llama-3.2-3B-Butler
Text Generation • Updated -
akhauriyash/Llama-3.2-1B-Butler
Text Generation • 1B • Updated -
TokenButler: Token Importance is Predictable
Paper • 2503.07518 • Published • 1