Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models.
Xianjun YangXiao WangQi ZhangLinda R. PetzoldWilliam Yang WangXun ZhaoDahua LinPublished in: CoRR (2023)
Keyphrases
- language model
- language modeling
- n gram
- speech recognition
- probabilistic model
- document retrieval
- query expansion
- information retrieval
- context sensitive
- retrieval model
- query terms
- language modelling
- statistical language models
- test collection
- smoothing methods
- relevance model
- pseudo relevance feedback
- document ranking
- word level
- language models for information retrieval
- language model for information retrieval
- spoken term detection
- ad hoc information retrieval
- vector space model
- keywords