Beyond single-shot evaluation: Measuring LLM capabilities through collaborative iteration
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).