Length-Adaptive Distillation: Customizing Small Language Model for Dynamic Token Pruning.
Chang LiuChongyang TaoJianxin LiangJiazhan FengTao ShenQuzhe HuangDongyan ZhaoPublished in: EMNLP (Findings) (2023)
Keyphrases
- language model
- language modeling
- n gram
- probabilistic model
- context sensitive
- retrieval model
- document retrieval
- information retrieval
- language modelling
- query expansion
- speech recognition
- mixture model
- ad hoc information retrieval
- document length
- language model for information retrieval
- statistical language models
- smoothing methods
- relevance model
- vector space model
- query terms
- test collection
- statistical machine translation
- pseudo relevance feedback
- word clouds