Sam Marks

Karma: 4,632

Evaluating honesty and lie detection techniques on a diverse suite of dishonest models

RowanWang, Sam Marks, Johannes Treutlein, evhub and Fabien Roger

25 Nov 2025 19:33 UTC

40 points

0 comments4 min readLW link

(alignment.anthropic.com)

Steering Evaluation-Aware Models to Act Like They Are Deployed

Tim Hua, andrq, Sam Marks and Neel Nanda

30 Oct 2025 15:03 UTC

61 points

12 comments16 min readLW link

Inoculation prompting: Instructing models to misbehave at train-time can improve run-time behavior

Sam Marks, Nevan Wichers, Daniel Tan, Aram Ebtekar, Jozdien, David Africa, Alex Mallen and Fabien Roger

8 Oct 2025 22:02 UTC

156 points

37 comments2 min readLW link

Petri: An open-source auditing tool to accelerate AI safety research

Sam Marks7 Oct 2025 20:39 UTC

77 points

0 comments1 min readLW link

(alignment.anthropic.com)

Eliciting secret knowledge from language models

Bartosz Cywiński, Arthur Conmy and Sam Marks

2 Oct 2025 20:57 UTC

68 points

3 comments2 min readLW link

(arxiv.org)

Discovering Backdoor Triggers

andrq, Tim Hua, Sam Marks, Arthur Conmy and Neel Nanda

19 Aug 2025 6:24 UTC

57 points

4 comments13 min readLW link

Towards Alignment Auditing as a Numbers-Go-Up Science

Sam Marks4 Aug 2025 22:30 UTC

127 points

15 comments6 min readLW link

Building and evaluating alignment auditing agents

Sam Marks, trentbrick, RowanWang, Sam Bowman, Euan Ong, Johannes Treutlein and evhub

24 Jul 2025 19:22 UTC

47 points

1 comment5 min readLW link

Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning

kh4dien, Helena Casademunt, Adam Karvonen, Sam Marks, Senthooran Rajamanoharan and Neel Nanda

23 Jul 2025 14:57 UTC

78 points

8 comments5 min readLW link

Principles for Picking Practical Interpretability Projects

Sam Marks15 Jul 2025 17:38 UTC

32 points

0 comments13 min readLW link

Race and Gender Bias As An Example of Unfaithful Chain of Thought in the Wild

Adam Karvonen and Sam Marks

2 Jul 2025 16:35 UTC

183 points

25 comments4 min readLW link

Unsupervised Elicitation of Language Models

Jiaxin Wen, Peter Hase, Sam Marks, Collin, Ethan Perez and janleike

13 Jun 2025 16:15 UTC

57 points

11 comments2 min readLW link

Modifying LLM Beliefs with Synthetic Document Finetuning

RowanWang, Johannes Treutlein, Avery, Ethan Perez, Fabien Roger and Sam Marks

24 Apr 2025 21:15 UTC

70 points

12 comments2 min readLW link

(alignment.anthropic.com)

Downstream applications as validation of interpretability progress

Sam Marks31 Mar 2025 1:35 UTC

112 points

3 comments7 min readLW link

Auditing language models for hidden objectives

Sam Marks, Johannes Treutlein, dmz, Sam Bowman, Hoagy, Carson Denison, Kei Nishimura-Gasparian, 7vik, Akbir Khan, Austin Meek, Euan Ong, Christopher Olah, Fabien Roger, jeanne_, Meg, Drake Thomas, Adam Jermyn, Monte M and evhub

13 Mar 2025 19:18 UTC

142 points

15 comments13 min readLW link

Recommendations for Technical AI Safety Research Directions

Sam Marks10 Jan 2025 19:34 UTC

64 points

1 comment17 min readLW link

(alignment.anthropic.com)

Alignment Faking in Large Language Models

ryan_greenblatt, evhub, Carson Denison, Benjamin Wright, Fabien Roger, Monte M, Sam Marks, Johannes Treutlein, Sam Bowman and Buck

18 Dec 2024 17:19 UTC

490 points

75 comments10 min readLW link

SAEBench: A Comprehensive Benchmark for Sparse Autoencoders

Can, Adam Karvonen, Johnny Lin, Curt Tigges, Joseph Bloom, chanind, Yeu-Tong Lau, Eoin Farrell, Arthur Conmy, CallumMcDougall, Kola Ayonrinde, Matthew Wearden, Sam Marks and Neel Nanda

11 Dec 2024 6:30 UTC

82 points

6 comments2 min readLW link

(www.neuronpedia.org)

Evaluating Sparse Autoencoders with Board Game Models

Adam Karvonen, Sam Marks, Can, Benjamin Wright, Jannik Brinkmann, Logan Riggs and Rico Angell

2 Aug 2024 19:50 UTC

38 points

1 comment9 min readLW link

Discriminating Behaviorally Identical Classifiers: a model problem for applying interpretability to scalable oversight

Sam Marks18 Apr 2024 16:17 UTC

115 points

10 comments12 min readLW link