home / rubrics / rubric_calibration_events

rubric_calibration_events: 38

This data as json

id teacher_id gradation_id_a gradation_id_b teacher_choice correct response_time_ms created_at confidence perceived_difficulty influential_feature margin rubric_version reasoning
38 deepseek-v4-pro 253 254 B 1 11771 2026-05-26 02:37:14     evidence_quality clearly 1 Response B includes a specific citation and data (elephants having less than 1% of natural range) and acknowledges a counterargument, while Response A uses only broad, unsupported claims.
Powered by Datasette · Queries took 0.631ms