LawrenceC

Karma: 5,366

I do AI Alignment research. Currently at METR, but previously at: Redwood Research, UC Berkeley, Good Judgment Project.

I’m also a part-time fund manager for the LTFF.

Obligatory research billboard website: https://chanlawrence.me/

OpenAI/Microsoft announce “next generation language model” integrated into Bing/Edge

LawrenceCFeb 7, 2023, 8:38 PM

79 points

4 comments1 min readLW link

(blogs.microsoft.com)

Evaluations (of new AI Safety researchers) can be noisy

LawrenceCFeb 5, 2023, 4:15 AM

132 points

11 comments16 min readLW link 1 review

The Alignment Problem from a Deep Learning Perspective (major rewrite)

SoerenMind, Richard_Ngo and LawrenceC

Jan 10, 2023, 4:06 PM

84 points

8 comments39 min readLW link

(arxiv.org)

Paper: Superposition, Memorization, and Double Descent (Anthropic)

LawrenceCJan 5, 2023, 5:54 PM

53 points

11 comments1 min readLW link

(transformer-circuits.pub)

Touch reality as soon as possible (when doing machine learning research)

LawrenceCJan 3, 2023, 7:11 PM

117 points

9 comments8 min readLW link 1 review

Shard Theory in Nine Theses: a Distillation and Critical Appraisal

LawrenceCDec 19, 2022, 10:52 PM

150 points

30 comments18 min readLW link

Paper: Constitutional AI: Harmlessness from AI Feedback (Anthropic)

LawrenceCDec 16, 2022, 10:12 PM

68 points

11 comments1 min readLW link

(www.anthropic.com)

Paper: Transformers learn in-context by gradient descent

LawrenceCDec 16, 2022, 11:10 AM

28 points

11 comments2 min readLW link

(arxiv.org)

Causal scrubbing: results on induction heads

LawrenceC, Adrià Garriga-alonso, Nicholas Goldowsky-Dill, ryan_greenblatt, Tao Lin, jenny, Ansh Radhakrishnan, Buck and Nate Thomas

Dec 3, 2022, 12:59 AM

34 points

1 comment17 min readLW link

Causal scrubbing: results on a paren balance checker

LawrenceC, Adrià Garriga-alonso, Nicholas Goldowsky-Dill, ryan_greenblatt, Tao Lin, jenny, Ansh Radhakrishnan, Buck and Nate Thomas

Dec 3, 2022, 12:59 AM

34 points

2 comments30 min readLW link

Causal scrubbing: Appendix

LawrenceC, Adrià Garriga-alonso, Nicholas Goldowsky-Dill, ryan_greenblatt, jenny, Ansh Radhakrishnan, Buck and Nate Thomas

Dec 3, 2022, 12:58 AM

18 points

4 comments20 min readLW link

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

LawrenceC, Adrià Garriga-alonso, Nicholas Goldowsky-Dill, ryan_greenblatt, jenny, Ansh Radhakrishnan, Buck and Nate Thomas

Dec 3, 2022, 12:58 AM

206 points

35 comments20 min readLW link 1 review

Paper: In-context Reinforcement Learning with Algorithm Distillation [Deepmind]

LawrenceCOct 26, 2022, 6:45 PM

29 points

5 comments1 min readLW link

(arxiv.org)

LawrenceC’s Shortform

LawrenceCOct 8, 2022, 5:17 PM

5 points

18 comments LW link

Paper: Discovering novel algorithms with AlphaTensor [Deepmind]

LawrenceCOct 5, 2022, 4:20 PM

82 points

18 comments1 min readLW link

(www.deepmind.com)

Language models seem to be much better than humans at next-token prediction

Buck, Fabien Roger and LawrenceC

Aug 11, 2022, 5:45 PM

182 points

60 comments13 min readLW link 1 review

High-stakes alignment via adversarial training [Redwood Research report]

dmz, LawrenceC and Nate Thomas

May 5, 2022, 12:59 AM

142 points

29 comments9 min readLW link

Book Review: Discrete Mathematics and Its Applications (MIRI Course List)

LawrenceCApr 14, 2015, 9:08 AM

21 points

12 comments8 min readLW link