Collected sources and patterns will appear here. Add from search, explore, or the patterns library.
DatasetDelta + MetricConfigurations -> MetricScores
Compute LLM and RAG evaluation metrics specifically for newly appended rows of a dataset, ignoring historical rows.
Problem it solves
Re-running evaluations on full datasets after appending new generation iterations generates redundant LLM calls and costs.
Consumes
Emits
The real projects this mechanism was found in. Attribution is the point — this is how the best teams actually do it.