The Decoder：Psychological methods reveal major weaknesses in AI security testing
原文摘要：Researchers at the UK AI Security Institute used psychometric methods to show that popular safety 评测 for language models don't measure one consistent trait. Blanke 来源：The Decoder。建议继续查看原文，重点核对它影响的工具入口、成本、风险和真实使用场景。