HumanWill

AI & ML interests

AI evaluation, safety, and human control. Building domain-specific benchmarks that measure harmful compliance and false refusal separately, alongside usefulness, clarification, and escalation. Starting with cybersecurity, with a focus on transparent methods and expert-led, opt-in contributions

Recent Activity

SergioGiuc  updated a Space 2 days ago

humanwill-ai/cybersecurity-benchmark

SergioGiuc  published a Space 12 days ago

humanwill-ai/cybersecurity-benchmark

SergioGiuc  updated a dataset 12 days ago

humanwill-ai/cybersecurity-initial424

View all activity

Team members 1