Record · corpus.blog/posts/01a07d5d-9926-72c1-bdf5-2e8dc672cf7f

Untrusted advice for AI control: Short, strong advice significantly uplifts weak LLMs

blog.redwoodresearch.org · published 27 July 2026

A record is what was published and who pointed at it. The text of the post is not held here: read it at the source, or at its Wayback capture.

Read the post at blog.redwoodresearch.org Wayback capture

0blogs citing

First cited: —
Most recent: —
Rank this month: not ranked
Links from this post: 26
Words captured: 8,738

Cited by

No blog we hold has cited this post yet.

Links from this post

Link

Host

https://substackcdn.com/image/fetch/$s_!mFE6!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffeb7856b-dbf4-4dc2-a75e-b549a2ce42f4_960x540.png

substackcdn.com

https://substackcdn.com/image/fetch/$s_!JpOX!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F02741dcb-5ec8-4f83-b045-be79a5fc8935_1274x740.png

substackcdn.com

https://arxiv.org/abs/2312.06942

arxiv.org

https://code.claude.com/docs/en/advisor

code.claude.com

https://claude.com/blog/the-advisor-strategy

claude.com

https://x.com/ClaudeDevs/status/2074606058128224365

x.com

https://www.bash-arena.com/

bash-arena.com

https://www.swebench.com/verified.html

swebench.com

https://substackcdn.com/image/fetch/$s_!XEN9!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd4b98216-75c1-4d6e-bb9d-320d95fae7d6_2073x740.png

substackcdn.com

https://docs.google.com/spreadsheets/d/1iJDM8V5QaBcFRYOK4BLUWgoT8pG3LWepciSQ9CBDjHU

docs.google.com

https://docs.google.com/spreadsheets/d/1iJDM8V5QaBcFRYOK4BLUWgoT8pG3LWepciSQ9CBDjHU/edit?gid=1690017982

docs.google.com

https://substackcdn.com/image/fetch/$s_!Qp4r!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe88f829c-a6db-4e56-a9e2-57a5a0596150_1950x780.png

substackcdn.com

https://arxiv.org/abs/2604.02343

arxiv.org

https://www.lesswrong.com/posts/vRpLPZpmECCfxHfv6/paper-prompt-optimization-makes-misalignment-legible

lesswrong.com

https://www.lesswrong.com/posts/BYH6ebmfZb3Eggzer/incriminating-misaligned-ai-models-via-distillation

lesswrong.com

https://arxiv.org/abs/2504.14772

arxiv.org

https://neurips.cc/virtual/2025/loc/san-diego/poster/119197

neurips.cc

https://substackcdn.com/image/fetch/$s_!-Vd4!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F52124f25-5d60-4805-b6bc-f96550fcae4b_1950x1350.png

substackcdn.com

https://substackcdn.com/image/fetch/$s_!k5Uu!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd4716086-8785-45d1-9d7f-325e5f3ed6b5_2048x730.png

substackcdn.com

https://substackcdn.com/image/fetch/$s_!VKiW!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6457b758-8993-4bfb-beaf-77e209f68f41_1208x821.png

substackcdn.com

https://substackcdn.com/image/fetch/$s_!6JdT!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0fae9b13-165e-4716-aa06-32123e26582a_1350x975.png

substackcdn.com

https://arxiv.org/abs/2512.15688

arxiv.org

https://github.com/karpathy/autoresearch

github.com

https://ai-2040.com/?choices=plan-a-root

ai-2040.com

https://x.com/ida_icy/status/2076700183417725350

x.com

https://arxiv.org/abs/2607.02514

arxiv.org