AK(@_akhaliq)
Thinking Inertia LLMs Keep Thinking When Told Not To paper: https://t.co/saG67wQ8Jw
6.4内容质量
TL;DR · AI 摘要
Title: AK @ akhaliq on X URL Source: Published Time: 2026-10-14T16:54:28.000Z Markdown Content: AK on X: "Thinking Inert...
核心要点
- 主题聚焦：Thinking Inertia LLMs Keep Thinking When Told No
- 来源：AK(@_akhaliq)，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AK on X: "Thinking Inertia LLMs Keep Thinking When Told Not To paper: https://t.co/saG67wQ8Jw" / X
[](https://x.com/)
AK on X: "Thinking Inertia LLMs Keep Thinking When Told Not To paper: https://t.co/saG67wQ8Jw"
-  AK @_akhaliq Thinking Inertia LLMs Keep Thinking When Told Not To paper: huggingface.co/papers/2610.11…
[](https://x.com/_akhaliq/status/2110413943034491358/photo/1) 4:54 PM · Oct 14, 2026·16.4K Views 20 11 79 60
-  Van0SS @Van0SS Oct 14 humans also think when they are told not to [](https://x.com/compose/post?in_reply_to=2110414608960049468) 1 117
-  Dream On ✨ | AI tools for dreamers@DR33MON Oct 14 Plain-English version: you tell the AI "don't overthink this" and it overthinks anyway. Finally, a model I relate to on a personal level 😅 Curious whether it's just extra tokens or actually changes the answers. [](https://x.com/compose/post?in_reply_to=2110414566714937631) 93
-  Asmir @asmirkn Oct 14 don't trust the no-think setting, just test it. run a fixed set of prompts with it on and off, log the output token counts, and flag it when the gap between them closes [](https://x.com/compose/post?in_reply_to=2110421076681314656) 29
See all the replies
Log in or sign up for X
See what’s happening and join the conversation
Continue with phoneContinue with Apple
Continue with Google
通过 Google 继续操作 通过 Google 继续操作。在新标签页中打开
or