Found 8 repositories(showing 8)
microsoft
A benchmark for evaluating the robustness of LLMs and defenses to indirect prompt injection attacks.
sam1gman
Automated LLM Threat Modeling & Red-Teaming Framework based on BIPIA v4.1. Evaluates prompt injection resilience with advanced obfuscation and defense matrix scoring.
chensishen
No description available
danielpiyo
No description available
No description available
Finetune LLMs with special <data></data> boundary tokens to defend against indirect prompt injection attacks (BIPIA benchmark)
No description available
No description available
All 8 repositories loaded