deepseek/deepseek-r1
20 runs · 20 questions
Aggregate
| Search | Scored / Total | Mean Cramér | Med |bias| | Output tokens |
|---|---|---|---|---|
| off | 20 / 20 | 0.4738 | 0.544 | 101,641 |
Cramér-log distribution
One bar per Cramér-log bucket. Buckets match the heatmap color scale: tight (<0.05), clean (<0.2), productive (<0.7), different-interpretation (<2), suspect (≥2).
Per-question Cramér-log
One row per question. Cells colored by Cramér-log band. Click a question to open its detail view.