Snowflake Model Explainability — Shapley-Value Feature Attribution (2026)

source.docs-snowflake-com.en-developer-guide-snowflake-ml-model-registry-model-explainability-2026 ↗

Snowflake documentation for Model Explainability: uses SHAP library to compute Shapley values attributing model predictions to input features. Supports XGBoost, CatBoost, LightGBM, Scikit-learn, and Snowpark ML classes. Currently in preview release (Snowpark ML 1.6.2+).