Lenny Rachitsky(@lennysan)
.@mlpowered shares one of the strangest model behaviors he's debugged at @AnthropicAI: "Slowly it w...
6.4内容质量
TL;DR · AI 摘要
Lenny Rachitsky on X: ".@mlpowered shares one of the strangest model behaviors he's debugged at @AnthropicAI: "Slowly it...
核心要点
- 主题聚焦：.@mlpowered shares one of the strangest model be
- 来源：Lenny Rachitsky(@lennysan)，建议结合原文判断细节。
- AI 分析暂不可用，本条为保底评分与摘要。
#AI
打开原文Lenny Rachitsky on X: ".@mlpowered shares one of the strangest model behaviors he's debugged at @AnthropicAI: "Slowly it would start descending into madness, writing in all caps, literally saying, 'I'm possessed by a demon." Full conversation: https://t.co/I3txx9wBvK" / X
Lenny Rachitsky
@lennysan
.
@
mlpowered
shares one of the strangest model behaviors he's debugged at
AnthropicAI
: "Slowly it would start descending into madness, writing in all caps, literally saying, 'I'm possessed by a demon." Full conversation:
youtu.be/JHo-aiePjoM
$
00:00
/$
Oct 18
"To make them is not to understand them." Emmanuel Ameisen (
) has spent the last 3 years doing interpretability research at
, looking inside LLMs like Claude to figure out what they're actually doing. As he puts it, models are "grown," not built. We
Show more
3:30 PM · Oct 19, 2026
·
5,578
Views
22
3