TokenButler -- Predict token importance for all heads across the transformer in the first layer itself. Enable fine-grained token sparsity!
YASH AKHAURI
akhauriyash
AI & ML interests
None yet
Recent Activity
authored
a paper
about 9 hours ago
ShadowLLM: Predictor-based Contextual Sparsity for Large Language Models
authored
a paper
about 9 hours ago
Attamba: Attending To Multi-Token States
authored
a paper
about 9 hours ago
TokenButler: Token Importance is Predictable
Organizations
None yet
Collections
1
models
5
akhauriyash/DeepSeek-R1-Distill-Llama-8B-Butler
Text Generation
•
Updated
•
16
akhauriyash/Llama-3.2-3B-Butler
Text Generation
•
Updated
•
10
akhauriyash/Llama-3.2-1B-Butler
Text Generation
•
Updated
•
44
akhauriyash/Llama-3.1-8B-Butler
Text Generation
•
Updated
•
10
akhauriyash/Llama-2-7b-hf-Butler
Text Generation
•
Updated
•
19
datasets
None public yet