Fine-mixing: Mitigating Backdoors in Fine-tuned Language Models.
Zhiyuan ZhangLingjuan LyuXingjun MaChenguang WangXu SunPublished in: CoRR (2022)
Keyphrases
- language model
- fine tuned
- fine tuning
- language modeling
- probabilistic model
- n gram
- domain specific
- speech recognition
- information retrieval
- document retrieval
- language modelling
- query expansion
- ad hoc information retrieval
- retrieval model
- context sensitive
- query terms
- test collection
- smoothing methods
- statistical language models
- pseudo relevance feedback
- document length
- language model for information retrieval
- term dependencies
- error rate
- okapi bm
- vector space model