SuffixDecoding: Extreme Speculative Decoding for Emerging AI Applications
NeurIPSSpotlight2025
TL;DR
Speculative decoding is widely adopted to reduce latency in large language model (LLM) inference by leveraging smaller draft models capable of handling diverse user tasks…
Opening excerpt from the authors’ abstract. source
Read the paper
Topics
large language model language model llm rag
← All NeurIPS 2025 Spotlight papers · Browse the whole archive