gpt-oss-safeguard
EvalsTool (not an eval — tooling/reference entry)
Open-weight safety reasoning classifiers (20B, 120B) driven by your policy.
- Category
- Guardrails & safety classifiers (adjacent)
- Insurance exposure
- Regulatory/bias
- Publisher
- OpenAI
- Publisher type
- Frontier lab
- License
- Apache-2.0
- Status
- Maintained (<12mo)
- Last commit
- 2026-01-14
- Stars
- 61
Wanich Risk Lab, 2026.