Alek Westover

Karma: 519

How to reduce capability degradation from off-model SFT

Dylan Xu, SebastianP and Alek Westover

8 Jun 2026 16:24 UTC

21 points

0 comments3 min readLW link

Advice for making robust-to-training model organisms

SebastianP, Alek Westover, Vivek Hebbar, Julian Stastny and Dylan Xu

28 May 2026 17:26 UTC

36 points

8 comments12 min readLW link

(blog.redwoodresearch.org)

Why does off-model SFT degrade capabilities?

SebastianP, Dylan Xu, Alek Westover, Julian Stastny and Vivek Hebbar

21 May 2026 0:35 UTC

40 points

9 comments6 min readLW link

Incriminating misaligned AI models via distillation

Alek Westover, SebastianP, Alex Mallen, Jozdien, Alexa Pan, Julian Stastny and Vivek Hebbar

15 May 2026 21:43 UTC

115 points

12 comments5 min readLW link

Sleeper Agent Backdoor Results Are Messy

SebastianP, Alek Westover, Dylan Xu, Vivek Hebbar and Julian Stastny

28 Apr 2026 1:55 UTC

81 points

4 comments7 min readLW link

An Empirical Study of Methods for SFTing Opaque Reasoning Models

SebastianP, Alek Westover, Vivek Hebbar, Dylan Xu and Julian Stastny

24 Apr 2026 17:26 UTC

17 points

0 comments6 min readLW link

How do LLMs generalize when we do training that is intuitively compatible with two off-distribution behaviors?

Dylan Xu, Alek Westover, Vivek Hebbar, SebastianP, frisby and Julian Stastny

20 Apr 2026 16:58 UTC

61 points

5 comments20 min readLW link

Five approaches to evaluating training-based control measures

Alek Westover, SebastianP, Julian Stastny and Vivek Hebbar

18 Apr 2026 1:07 UTC

21 points

0 comments6 min readLW link

Model organisms researchers should check whether high LRs defeat their model organisms

Dylan Xu, SebastianP, Alek Westover, Vivek Hebbar and Julian Stastny

10 Apr 2026 0:07 UTC

40 points

0 comments5 min readLW link

How will we do SFT on models with opaque reasoning?

Alek Westover, Vivek Hebbar and egan

21 Feb 2026 0:00 UTC

32 points

17 comments7 min readLW link

Three visions for diffuse control

Alek Westover9 Feb 2026 6:41 UTC

8 points

0 comments3 min readLW link

Theoretical predictions on the sample efficiency of training policies and activation monitors

Alek Westover and Vivek Hebbar

10 Jan 2026 23:50 UTC

18 points

2 comments7 min readLW link

Four Downsides of Training Policies Online

Alek Westover and egan

4 Jan 2026 3:17 UTC

29 points

4 comments3 min readLW link

Methodological considerations in making malign initializations for control research

Alek Westover, Vivek Hebbar and Julian Stastny

24 Dec 2025 1:18 UTC

16 points

0 comments13 min readLW link

Notes on Software-Based Compute-Usage Verification

Alek Westover15 Dec 2025 3:40 UTC

9 points

0 comments12 min readLW link

Alek Westover’s Shortform

Alek Westover8 Dec 2025 4:24 UTC

2 points

20 comments1 min readLW link

Should AI Developers Remove Discussion of AI Misalignment from AI Training Data?

Alek Westover23 Oct 2025 15:12 UTC

51 points

3 comments9 min readLW link

What training data should developers filter to reduce risk from misaligned AI? An initial narrow proposal

Alek Westover17 Sep 2025 15:30 UTC

44 points

4 comments18 min readLW link

Why I think AI will go poorly for humanity

Alek Westover19 Mar 2025 15:52 UTC

14 points

0 comments30 min readLW link

Safe Distillation With a Powerful Untrusted AI

Alek Westover20 Feb 2025 3:14 UTC

5 points

1 comment5 min readLW link