[2604.23483] Agentic Adversarial Rewriting Exposes Architectural Vulnerabilities in Black-Box NLP Pipelines
![[2604.23483] Agentic Adversarial Rewriting Exposes Architectural Vulnerabilities in Black-Box NLP Pipelines](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
Summary
Abstract page for arXiv paper 2604.23483: Agentic Adversarial Rewriting Exposes Architectural Vulnerabilities in Black-Box NLP Pipelines
Description
Abstract page for arXiv paper 2604.23483: Agentic Adversarial Rewriting Exposes Architectural Vulnerabilities in Black-Box NLP Pipelines
Original reporting
AFBytes is a read-only aggregator. Use the original source for full context and complete reporting.
Open original source