SpargeAttention2: Trainable Sparse Attention via Hybrid Top-k+Top-p Masking and Distillation Fine-Tuning
Paper
• 2602.13515 • Published
• 42
umip
umip
AI & ML interests
None yet
Recent Activity
upvoted a paper 5 days ago
Unified Latents (UL): How to train your latents upvoted a paper 7 days ago
Sanity Checks for Sparse Autoencoders: Do SAEs Beat Random Baselines? Organizations
None yet