sk Skill · hparreao
confidence-calibration
Audit confidence estimates against observed LLM or agent outcomes. Use when routing by confidence, comparing logprob or ensemble confidence, monitoring calibration drift, or validating uncertainty claims.
Open on skills.sh ↗read 2026-09-19
- installs 8w
- 0
- 30-day movement
- starts with the next reading
- Related entries
- 2
- Connections
- 0
Python
- Host repository
- hparreao/Awesome-AI-Evaluation-Guide
- Host stars
- 19
- Host language
- Python