TrojLLM: A Black-box Trojan Prompt Attack on Large Language Models.
Jiaqi XueMengxin ZhengTing HuaYilin ShenYepeng LiuLadislau BölöniQian LouPublished in: NeurIPS (2023)
Keyphrases
- black box
- language model
- language modeling
- probabilistic model
- n gram
- retrieval model
- black boxes
- white box
- information retrieval
- document retrieval
- statistical language models
- test collection
- language modelling
- query expansion
- speech recognition
- ad hoc information retrieval
- query terms
- context sensitive
- document ranking
- pseudo relevance feedback
- integration testing
- translation model
- white box testing
- statistical language modeling
- decision trees