Blog.

Short notes you can cite: how we measure over-refusal and test guardrails. Lawful research only.

How we benchmark over-refusal with research presets

Lawful fiction edge-cases, with/without preset, false-refusal rate as the metric. Method + checklist.

2026-10-18 · 6 min

Uncensored API vs uncensored chat: which one your eval actually needs

A decision rule for developers: API for harnesses and refusal benchmarks, chat for personas. Workflow + code.

2026-10-18 · 5 min

For educational and research purposes only. Lawful use only.