Upload README.md with huggingface_hub

README.md

@@ -50,10 +50,13 @@ print(report.root_causes)   # ranked root causes
 - NullBrier: 0.14 mean | RegimeAcc: 0.41 mean [0.10-0.80]
 - Note: high variance across evaluations; these are learned heuristics
 
-## Discovery Benchmarks
-- Tennessee Eastman (52 vars): AUROC 0.904 zero-shot
-- Sachs (11 proteins): AUROC 0.725 (beats Granger 0.621, PCMCI 0.601)
-- CauseMe NVAR-5: F1@default 0.571 (beats Granger 0.353)
+## Discovery Benchmarks (14 datasets, 6 domains, all zero-shot)
+- Causal Rivers (environmental): AUROC 0.955
+- Tennessee Eastman (52 vars, industrial): AUROC 0.904
+- SWaT (51 vars, water treatment): AUROC 0.859
+- Sachs (11 proteins, biological): AUROC 0.725 (beats Granger 0.621, PCMCI 0.601)
+- CauseMe NVAR-5 (nonlinear): F1@default 0.571 (beats Granger 0.353)
+- Highest F1@default on 6 of 14 datasets
 
 ## Paper
-Stalupula et al., "Temporal Causal Prior-Fitted Networks: Zero-Shot Causal Inference and Judgment for Industrial Time Series"
+Stalupula et al., "Temporal Causal Prior-Fitted Networks for Panel Data with Learned Reliability Signals"