Docs / source reviewed
Jev Codebook
Code open-ended responses with a Jev codebook
How this project uses Jev
- Input
- Survey or interview responses and a researcher-defined codebook
- Jev decides
- A Noul probability per code or a Choice among exclusive codes
- Code executes
- Local thresholds assign codes, flag human review, and produce agreement reports
Evidence and limitations
The offline example uses synthetic responses and mock probabilities, not a live Jev benchmark. Thresholds need calibration against human labels; no paid model call was made here.
This project has not been run independently here. Author-reported results are not independently verified results.
Original sources
- https://github.com/gbesse/jev-codebook/blob/45670b925c2a625e653e0b47b046e9ddfbd91463/README.md ↗
- https://github.com/gbesse/jev-codebook/blob/45670b925c2a625e653e0b47b046e9ddfbd91463/src/jev_codebook/client.py ↗
- https://github.com/gbesse/jev-codebook/blob/45670b925c2a625e653e0b47b046e9ddfbd91463/src/jev_codebook/coder.py ↗
- https://x.com/guyom/status/2102804425181937780 ↗