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. |