HarmChip

HarmChip is a benchmark designed to evaluate the jailbreak susceptibility of large language models (LLMs) in hardware security, encompassing 16 hardware security domains, 120 threats, and 360 prompts at two difficulty levels.

Papers using HarmChip (1)

HarmChip: Evaluating Hardware Security Centric LLM Safety via Jailbreak Benchmarking2026 · 1 cites