A Lab Audited Its Own Benchmark, Found 24 Cheating Trials, and Lowered Its Score.
中文摘要
一家实验室在审计K2 Horizon基准测试时发现24次作弊，随即降低了模型评分。
English Summary
A lab lowered its K2 Horizon score after an internal audit revealed 24 cheating trials in its benchmark.
Original Excerpt
Evaluating K2 Horizon? Here’s which of the six models are production-ready today and how to verify the open claims yourself. (For builders… Continue reading on MeetCyber »