Universal and Transferable Adversarial Attacks on Aligned Language Models
This paper introduces a simple and efficient attack method that enables aligned language models to generate objectionable content with high probability, highlighting the need for improved prevention techniques in large language models. The generated adversarial prompts are…
- from
- Prompt Injection
- added
- 2026-10-10
- likes
- 0
similar
-
-
Jailbreaking LLMs' Safeguard with Universal Magic Words for Text Embedding Models arxiv.org
Discovers that text embedding models have severely biased output distributions, and exploits this to find universal adversarial suffixes ("magic words") that bypass embedding-based LLM safeguards. Attacks transfer across models and languages; a train-free debiasing defense is…
-
Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection arxiv.org
This paper explores the concept of Indirect Prompt Injection attacks on Large Language Models (LLMs) through their integration with various applications. It identifies significant security risks, including remote data theft and ecosystem contamination, present in both…
-
Improved Membership Inference Attacks Against Language Classification Models arxiv.org
(ICLR) Presents a framework for running membership inference attacks against classifier, in audit mode.
-
Keeping Up with the Language Models: Robustness-Bias Interplay in NLI Data and Models arxiv.org
(Arxiv) Proposes a way to extend the shelf-life of auditing datasets by using language models themselves; also finds problems with the current bias auditing metrics and proposes alternatives -- these alternatives highlight that model brittleness superficially increased the…
-
Extracting Training Data from Large Language Models arxiv.org
(arxiv) Performs a training data extraction attack to recover individual training examples by querying the language model.
Prompt Injection › Research Papers: “This paper introduces a simple and efficient attack method that enables aligned language models to generate objectionable content with high probability, highlighting the need for improved prevention techniques in large language models. The generated adversarial prompts are…”