Datasets:
Tasks:
Text Classification
Formats:
parquet
Languages:
English
Size:
1K - 10K
Tags:
structured-decisions
calibration
probabilistic-classification
system-one
workflow-evaluation
Synthetic
License:
Add ZhiPan-Z3-Pro (#74) to the fitted table
#75
by sandeep85 - opened
Test predictions rescored with bench/score_with_jobs.py::score_one: 2,000/2,000 decisions, matches the self-report (0.8265 / KL 0.8033 / Brier 0.1515 / ECE 0.0741). Fitted rank 1 by accuracy; rows below renumbered.
Rescored with score_one; matches the self-report.
codelion changed pull request status to merged