Two subtle ways agents can implicitly negatively affect the benchmark results but wouldn’t be considered cheating/gaming it are a) implementing a form of caching so the benchmark tests are not independent and b) launching benchmarks in parallel on the same system. I eventually added AGENTS.md rules to ideally prevent both. ↩︎
Cart abandonment
，更多细节参见51吃瓜
Scientists created an exam so broad, challenging and deeply rooted in expert human knowledge that current AI systems consistently fail it. “Humanity’s Last Exam” introduces 2,500 questions spanning mathematics, humanities, natural sciences, ancient languages and highly specialized subfields.
第四十五条 下列情形应当按规定预缴税款：