S2D: Sorted Speculative Decoding For More Efficient Deployment of Nested Large Language Models.
Parsa KavehzadehMohammadreza PourrezaMojtaba ValipourTinashu ZhuHaoli BaiAli GhodsiBoxing ChenMehdi RezagholizadehPublished in: CoRR (2024)
Keyphrases
- language model
- language modeling
- probabilistic model
- document retrieval
- n gram
- statistical language models
- retrieval model
- language modelling
- language models for information retrieval
- query expansion
- smoothing methods
- test collection
- information retrieval
- query terms
- language model for information retrieval
- document ranking
- pseudo relevance feedback
- vector space model
- context sensitive
- speech recognition
- document length
- information retrieval systems
- active learning