Survey of Vulnerabilities in Large Language Models Revealed by Adversarial Attacks.
Erfan ShayeganiMd Abdullah Al MamunYu FuPedram ZareeYue DongNael B. Abu-GhazalehPublished in: CoRR (2023)
Keyphrases
- language model
- countermeasures
- security vulnerabilities
- security flaws
- security risks
- language modeling
- security mechanisms
- document retrieval
- speech recognition
- probabilistic model
- buffer overflow
- query expansion
- data corruption
- information security
- information retrieval
- statistical language models
- n gram
- language modelling
- smoothing methods
- retrieval model
- query terms
- context sensitive
- attack scenarios
- pseudo relevance feedback
- word error rate
- language models for information retrieval
- security issues
- dos attacks
- relevance model
- vector space model
- web search