MEASURING LLM SIMILARITY

Summary from the TLDR AI newsletter

CAPA is a metric that assesses model similarity based on shared errors.