[2606.03819] TreeFlash: Parallel AR-Approximation for Faster Speculative Decoding
![[2606.03819] TreeFlash: Parallel AR-Approximation for Faster Speculative Decoding](/static/img/curated/desks/afbytes-us.png)
AI disclosure
Summary
Abstract page for arXiv paper 2606.03819: TreeFlash: Parallel AR-Approximation for Faster Speculative Decoding