Collapse Index (CI) & Structural Retention Index (SRI): Advanced diagnostic frameworks for bounded, lightweight evaluation of AI model instability beyond accuracy, revealing hidden brittleness in production systems!

Public validation of SRI (Structural Retention) and CI (Collapse Index) metrics on AG News dataset: 90.4% accuracy, 9.2% flip rate, revealing model brittleness beyond standard benchmarks with AUC 0.874!

Public validation of Collapse Index (CI) on SST-2 dataset: 42.8% flip rate, AUC 0.698. Reveals model brittleness beyond 90%+ accuracy under perturbations!

CI-Guided Data Curation: Using prediction instability to detect label noise. Validated on SST-2 with 10% noise injection. Recovered 40% of accuracy gap to perfect labels. Part of the Collapse Index framework.

🔍 Analyze and evaluate datasets with Collapse Index for accurate insights and comprehensive reports, enhancing your research and decision-making process.