Upload results/benchmark_report.md with huggingface_hub
Browse files- results/benchmark_report.md +29 -0
results/benchmark_report.md
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# CausalFinBench Results
|
| 2 |
+
|
| 3 |
+
Date: 2026-05-08 00:12
|
| 4 |
+
Model: Horizon v1 (172M DiT, Pearl Level 3)
|
| 5 |
+
Checkpoint: checkpoints/phase9_production/step_200000.pt
|
| 6 |
+
Total time: 26.0 minutes
|
| 7 |
+
|
| 8 |
+
## Summary
|
| 9 |
+
|
| 10 |
+
- Tier A: 5/5 PASS
|
| 11 |
+
- Tier B: 2/3 PARTIAL
|
| 12 |
+
- Tier C: 1/1 PASS
|
| 13 |
+
|
| 14 |
+
**OVERALL: ISSUES FOUND**
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
## Detailed Results
|
| 18 |
+
|
| 19 |
+
| Test | Tier | Name | Cases | Pass Rate | Result |
|
| 20 |
+
|------|------|------|-------|-----------|--------|
|
| 21 |
+
| A1 | A | Consistency | 100 | 100.0% | PASS |
|
| 22 |
+
| A2 | A | Causal Asymmetry | 19 | 100.0% | PASS |
|
| 23 |
+
| A3 | A | Compositionality | 30 | 93.3% | PASS |
|
| 24 |
+
| A4 | A | Counterfactual Coherence | 50 | 92.0% | PASS |
|
| 25 |
+
| A5 | A | Robustness | 56 | 100.0% | PASS |
|
| 26 |
+
| B1 | B | Placebo (non-edges) | 39 | 59.0% | FAIL |
|
| 27 |
+
| B2 | B | Real effects (edges) | 19 | 100.0% | PASS |
|
| 28 |
+
| B4 | B | Sensitivity monotonicity | 10 | 100.0% | PASS |
|
| 29 |
+
| C1 | C | RBI Rate Decisions | 42 | 100.0% | PASS |
|