サインイン
インサイト - AI Evaluation
AI Evaluation and Red Teaming: Safe Harbor Proposal
Major AI developers should provide legal and technical safe harbors to protect public interest safety research from account suspensions or legal reprisal.
LatEval: Evaluating LLMs with Lateral Thinking Puzzles
LLMs struggle with lateral thinking in LatEval, highlighting the need for improved AI capabilities.
Automated Dynamic Evaluation of AI Assistants' API Invocation Capabilities
AutoDE provides a dynamic evaluation framework that closely mirrors human assessments, revealing deficiencies overlooked by static evaluations.