Docent
EvalsTool (not an eval — tooling/reference entry)
Analyze agent transcripts at scale for failures and reward hacking.
- Category
- AI control & monitoring
- Insurance exposure
- Model failure
- Publisher
- Transluce
- Publisher type
- Nonprofit / research org
- License
- Apache-2.0
- Status
- Active (<90d)
- Last commit
- 2026-10-08
- Stars
- 141
Wanich Risk Lab, 2026.