sterlixlol/kazi-agentic-traces-19k
Kazi Agentic SWE Traces — 19K 18,974 multi-turn agentic software-engineering traces in the OpenAI messages + tools schema, judge-filtered to an average score of ≥ 9.0/10. Built to fine-tune Kazi, an uncensored coding / sysadmin / research agent, on an abliterated Qwen3.6-27B base — released under ODC-BY for anyone building Kazi-class agents. What this is — and what it isn't ✅ Complete simulated agentic sessions. Each trace runs end-to-end: system prompt → user… See the full description on the dataset page: https://huggingface.co/datasets/sterlixlol/kazi-agentic-traces-19k.
De-dupe to 18,974 (drop 23 repetition traces)
Tighten card: explicit simulated-output disclosure, judge caveat, limitations, usage + example
Replace with PII-scrubbed version: names randomized per-record (70-name pool), /home/sterlix to /home/user, username/email/agency redacted
Add fine-tuning instructions
Use HTML table for headerless layout
Convert summary table to bullet list
Fix empty table headers
Add 18,997 judge-filtered traces
Add dataset card
initial commit