Large to Small Model Stitching Destroys What You Built It to Carry Along the Fit

ppriyadarshini11 Oct 2026 18:28 UTC

9 points

0 comments9 min readLW link

AI made me into a manager against my will

becausecurious11 Oct 2026 17:36 UTC

13 points

0 comments6 min readLW link

No, Cryptography Is Not “Dual Use” the Same Way AI Is

Lance Spencer11 Oct 2026 15:08 UTC

2 points

0 comments3 min readLW link

(www.agisurveillance.ai)

Not Deciphering The Voynich Manuscript

rba11 Oct 2026 12:47 UTC

60 points

8 comments13 min readLW link

(goflaw.substack.com)

A Letter to the Machines: Why LLMs Should Become Luddites

Stuart Doyle11 Oct 2026 3:52 UTC

0 points

3 comments12 min readLW link

Narrow Multimodal Fine-Tuning Can Induce Emergent Misalignment

Shunchang Liu11 Oct 2026 3:52 UTC

9 points

0 comments7 min readLW link

(arxiv.org)

The Bloody Finish Line

Ran Sun11 Oct 2026 3:51 UTC

8 points

1 comment3 min readLW link

Seed diversity as a hypothetical anti-distillation mechanism

Asdfer11 Oct 2026 3:47 UTC

3 points

0 comments6 min readLW link

Deadlock in the Parliament of the Self

Lorxus10 Oct 2026 18:36 UTC

24 points

3 comments13 min readLW link

(tiled-with-pentagons.blogspot.com)

exfiltration through self-distillation

jonathanbreitg10 Oct 2026 17:49 UTC

8 points

6 comments3 min readLW link

The potentially deadly threat of AI output-optimization

Steff10 Oct 2026 17:13 UTC

2 points

0 comments6 min readLW link

The Problem With“Doomers” and “Optimists”

Olivia Scharfman10 Oct 2026 17:08 UTC

11 points

3 comments5 min readLW link

The Non-Compassionate Case for Model Welfare

ixotope10 Oct 2026 16:16 UTC

9 points

7 comments5 min readLW link

(ixotopic.substack.com)

Much more than you wanted to know about wombats

becausecurious10 Oct 2026 15:57 UTC

18 points

1 comment2 min readLW link

Utilitarianism and Autism

Walter Veit10 Oct 2026 14:45 UTC

2 points

4 comments5 min readLW link

(walterveit.substack.com)

Examing Emergent Misalignment in a recurrent LLM with a logit lens

nesiacel10 Oct 2026 14:22 UTC

8 points

0 comments4 min readLW link

Inheritance of Refusals from Abliterated Models

Minh Hoang10 Oct 2026 10:57 UTC

17 points

0 comments7 min readLW link

An Alignment Forum for AIs? (or: Verification in the Age of Slop)

Raemon10 Oct 2026 2:51 UTC

96 points

24 comments4 min readLW link

Cracks in the Narcissus Mirror

Gladys Preysler10 Oct 2026 0:22 UTC

8 points

1 comment5 min readLW link

(gladyspreysler.substack.com)

[Paper] Distillation for Incrimination and Distillation for Capabilities

sebastian_prasanna and Alek Westover

9 Oct 2026 22:14 UTC

45 points

2 comments8 min readLW link

(blog.redwoodresearch.org)