arxiv:2504.04823
SunYuxuan
snowdusky
AI & ML interests
None yet
Recent Activity
upvoted a paper 2 months ago
Attention Sink in Transformers: A Survey on Utilization, Interpretation, and Mitigation upvoted a paper 2 months ago
TriAttention: Efficient Long Reasoning with Trigonometric KV CompressionOrganizations
None yet