RSS

Nicholas Goldowsky-Dill

Karma: 908

Interpretability Researcher at Apollo Research

Claude Son­net 3.7 (of­ten) knows when it’s in al­ign­ment evaluations

17 Mar 2025 19:11 UTC
184 points
9 comments6 min readLW link

De­tect­ing Strate­gic De­cep­tion Us­ing Lin­ear Probes

6 Feb 2025 15:46 UTC
104 points
9 comments2 min readLW link
(arxiv.org)

Ni­cholas Goldowsky-Dill’s Shortform

Nicholas Goldowsky-Dill6 Nov 2024 12:37 UTC
5 points
2 commentsLW link

A List of 45+ Mech In­terp Pro­ject Ideas from Apollo Re­search’s In­ter­pretabil­ity Team

18 Jul 2024 14:15 UTC
123 points
18 comments18 min readLW link

Apollo Re­search 1-year update

29 May 2024 17:44 UTC
93 points
0 comments7 min readLW link

The Lo­cal In­ter­ac­tion Ba­sis: Iden­ti­fy­ing Com­pu­ta­tion­ally-Rele­vant and Sparsely In­ter­act­ing Fea­tures in Neu­ral Networks

20 May 2024 17:53 UTC
108 points
4 comments3 min readLW link

Iden­ti­fy­ing Func­tion­ally Im­por­tant Fea­tures with End-to-End Sparse Dic­tionary Learning

17 May 2024 16:25 UTC
57 points
20 comments4 min readLW link
(arxiv.org)

Causal scrub­bing: re­sults on in­duc­tion heads

3 Dec 2022 0:59 UTC
34 points
1 comment17 min readLW link

Causal scrub­bing: re­sults on a paren bal­ance checker

3 Dec 2022 0:59 UTC
34 points
2 comments30 min readLW link

Causal scrub­bing: Appendix

3 Dec 2022 0:58 UTC
18 points
4 comments20 min readLW link

Causal Scrub­bing: a method for rigor­ously test­ing in­ter­pretabil­ity hy­pothe­ses [Red­wood Re­search]

3 Dec 2022 0:58 UTC
206 points
35 comments20 min readLW link1 review