Event ·Multi-source

AA-PersonalityBench评测:Claude Fable 5.2回答更直接、更简短

Artificial Analysis 发布的 AA-PersonalityBench 评测显示,Claude Fable 5.2 相比 Claude Fable 5.1 在 Direct(直接性)和 Answer First(先给答案)两项上排名更高,但在 Informativeness(信息量)和 Conviction(确定性)上有所下降。风格层面的变化比性格层面更明显,沟通特征平均移动了 38 个百分点。具体表现上,5.2 写出的回复明显更短,几乎不再使用破折号(em dash)。这份评测说明 Anthropic 在版本迭代中对输出风格做了较大调整,优先压缩冗余表达、直奔主题。

Current read

如果你嫌 5.1 说话绕、爱用破折号,这个评测值得看:5.2 回答更短更直接,但信息量和确定性反而降了一点。关心模型输出风格差异的开发者可以对照参考。

11 sources46° AI heatLast updated 2026/11/16 23:42:13

Evidence

11 sources
Primary sourceArtificial Analysis
View source
Related source 1The Rundown AI
View source
Related source 3kimmonismus
View source
Related source 6lmarena.ai
View source
Related source 7marktechpost
View source
Related source 8掘金本周最热
View source
Related source 9TestingCatalog
View source
Related source 10Yuchen Jin
View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

AA-PersonalityBench评测:Claude Fable 5.2回答更直接、更简短 · 事件专题 | AITOP