TokenButler
Collection
TokenButler -- Predict token importance for all heads across the transformer in the first layer itself. Enable fine-grained token sparsity!
•
6 items
•
Updated
•
2
YASH AKHAURI
akhauriyash
AI & ML interests
None yet
Recent Activity
authored
a paper
about 14 hours ago
ShadowLLM: Predictor-based Contextual Sparsity for Large Language Models
authored
a paper
about 14 hours ago
Attamba: Attending To Multi-Token States
authored
a paper
about 14 hours ago
TokenButler: Token Importance is Predictable
Organizations
None yet