Safeguards/Safety, NLP, Security, Text Classification, Transformers and Agents
Detect and redact prompt injection with span precision
