TurnTrout

Karma: 21,265

I don’t use LessWrong much anymore. Find me at www.turntrout.com.

My name is Alex Turner. I’m a research scientist at Google DeepMind on the Scalable Alignment team. My views are strictly my own; I do not represent Google. Reach me at alex[at]turntrout.com

Training a Reward Hacker Despite Perfect Labels

ariana_azarbal, vgillioz and TurnTrout

14 Aug 2025 23:57 UTC

127 points

45 comments4 min readLW link

Optimizing The Final Output Can Obfuscate CoT (Research Note)

lukemarks, jacob_drori, cloud and TurnTrout

30 Jul 2025 21:26 UTC

195 points

22 comments6 min readLW link

English writes numbers backwards

TurnTrout25 Jul 2025 23:00 UTC

8 points

23 comments12 min readLW link

(turntrout.com)

We Built a Tool to Protect Your Dataset From Simple Scrapers

TurnTrout, Edward Turner, Dipika Khullar and royrin

25 Jul 2025 5:44 UTC

55 points

9 comments3 min readLW link

A Simple Explanation of AGI Risk

TurnTrout1 Jul 2025 16:18 UTC

66 points

4 comments5 min readLW link

(turntrout.com)

Authors Have a Responsibility to Communicate Clearly

TurnTrout1 Jul 2025 15:41 UTC

125 points

29 comments6 min readLW link

(turntrout.com)

Distillation Robustifies Unlearning

Bruce W. Lee, Addie Foote, alexinf, leni, Jacob G-W, Harish Kamath, Bryce Woodworth, cloud and TurnTrout

13 Jun 2025 13:45 UTC

232 points

43 comments8 min readLW link

(arxiv.org)

Self-fulfilling misalignment data might be poisoning our AI models

TurnTrout2 Mar 2025 19:51 UTC

154 points

29 comments1 min readLW link

(turntrout.com)

Steering Gemini with BiDPO

TurnTrout31 Jan 2025 2:37 UTC

104 points

5 comments1 min readLW link

(turntrout.com)

Insights from “The Manga Guide to Physiology”

TurnTrout24 Jan 2025 5:18 UTC

26 points

3 comments1 min readLW link

(turntrout.com)

Deceptive Alignment and Homuncularity

Oliver Sourbut and TurnTrout

16 Jan 2025 13:55 UTC

26 points

12 comments22 min readLW link

Gaming TruthfulQA: Simple Heuristics Exposed Dataset Weaknesses

TurnTrout16 Jan 2025 2:14 UTC

64 points

3 comments1 min readLW link

(turntrout.com)

Review: Breaking Free with Dr. Stone

TurnTrout18 Dec 2024 1:26 UTC

47 points

5 comments1 min readLW link

(turntrout.com)

Gradient Routing: Masking Gradients to Localize Computation in Neural Networks

cloud, Jacob G-W, Evzen, Joseph Miller and TurnTrout

6 Dec 2024 22:19 UTC

169 points

12 comments11 min readLW link

(arxiv.org)

Deep Causal Transcoding: A Framework for Mechanistically Eliciting Latent Behaviors in Language Models

Andrew Mack and TurnTrout

3 Dec 2024 21:19 UTC

106 points

8 comments41 min readLW link

Intrinsic Power-Seeking: AI Might Seek Power for Power’s Sake

TurnTrout19 Nov 2024 18:36 UTC

40 points

5 comments1 min readLW link

(turntrout.com)

Announcing turntrout.com, my new digital home

TurnTrout17 Nov 2024 17:42 UTC

108 points

33 comments1 min readLW link

(turntrout.com)

I found >800 orthogonal “write code” steering vectors

Jacob G-W and TurnTrout

15 Jul 2024 19:06 UTC

104 points

19 comments7 min readLW link

(jacobgw.com)

Mechanistically Eliciting Latent Behaviors in Language Models

Andrew Mack and TurnTrout

30 Apr 2024 18:51 UTC

214 points

43 comments45 min readLW link

Many arguments for AI x-risk are wrong

TurnTrout5 Mar 2024 2:31 UTC

168 points

87 comments12 min readLW link