Science & Engineering
CausalDynamics
Causal discovery on thousands of coupled differential-equation dynamical systems.
Response matrix
Every model, scored item by item.
Each row is an AI model and each column an item, ordered so the strongest models and easiest items gather toward one corner. 10 subjects × 14,617 items, 94% of cells evaluated.
Fit to width. Hover for subject & item; click a cell for details.
Scale: Per metric (each shown on its own scale): AUROC / AUPRC in [0, 1] for edge recovery (higher is better); SHD = structural Hamming distance, an unbounded error count (lower is better).
Sample items
What the questions look like — and how subjects answer.
A spread of items across the difficulty range. This benchmark does not publish per-answer traces, so each item shows which subjects succeeded.
Subjects
The models, agents, and reward models evaluated.
10 subjects, ranked by mean response across this benchmark's items.
- 14f5a97d859.221
- 2215d7abc26.571
- 3be55b8a523.013
- 489e7831921.106
- 575d906b920.619
- 6b15b949319.204
- 771af175118.207
- 8688a9b3f17.318
- 960d8888a11.853
- 1090d0a19c9.617