Confidently wrong: AI fact-checkers show the Dunning-Kruger effect
中文摘要
小型AI模型在事实核查中表现出达克效应：虽然错误较多，却对错误答案表现出极高自信。
English Summary
Smaller AI models exhibit the Dunning-Kruger effect, making more mistakes while remaining highly confident in their incorrect answers.
原文节选
Smaller AI models make more mistakes but also think they’re right Continue reading on Medium »