Skip to content

Topic

LLM evaluation infrastructure

The systems used to run, replay and score language-model jobs outside the serving path, including canary and shadow execution tiers.

Current clusters