mirror of
https://github.com/pewdiepie-archdaemon/odysseus.git
synced 2026-10-06 15:02:20 +02:00
899 B
899 B
Historical Odysseus QA Queue
- Source sessions: 1294
- Source user turns / teacher seeds: 3258
- Unique conversation flows: 596
- Historical labels are conservative;
replay_firstmust be replayed before assigning ownership.
Workstreams
harness: 1model_sft: 1backend: 1replay_first: 593
Families
calendar: 421cookbook_admin: 203documents: 173email: 359general: 303memory: 179notes: 362research: 14search_browser: 459shell_files: 104skills: 226switching: 130tasks: 197ui: 128
Workflow
- Cook one fresh conversation from every seed using the complete tool catalog.
- Replay safe cooked cases on the current 7011 Agent runtime.
- Judge, classify ownership, and patch recurring behavior classes.
- Retain duplicate source runs as stability evidence; account for quarantined cases explicitly.