Signal2026-07-28
arXiv

When LLM Defenses Backfire: Characterizing Safety, Performance, and Cost Trade-offs

Part of

Adaptive Interfaces And AI Benchmarking

Open primary sourceSee the whole picture