Evaluation
Platform and SDK capabilities for assessing model performance across data splits, running error and slice analysis, and building expert-curated benchmarks. Largely delivered in-platform and through the SDK against a customer's Snorkel Flow instance rather than a public REST endpoint.