nostalgebraist

Karma: 5,615

Same person as nostalgebraist2point0, but now I have my account back.

Elsewhere:

I have signed no contracts or agreements whose existence I cannot mention.

the case for CoT unfaithfulness is overstated

nostalgebraist29 Sep 2024 22:07 UTC

242 points

40 comments11 min readLW link

instruction tuning and autoregressive distribution shift

nostalgebraist5 Sep 2024 16:53 UTC

40 points

5 comments5 min readLW link

[Question] Am I confused about the “malign universal prior” argument?

nostalgebraist27 Aug 2024 23:17 UTC

92 points

33 comments8 min readLW link

nostalgebraist’s Shortform

nostalgebraist27 May 2024 17:49 UTC

7 points

8 comments1 min readLW link

OpenAI API base models are not sycophantic, at any size

nostalgebraist29 Aug 2023 0:58 UTC

182 points

20 comments2 min readLW link

(colab.research.google.com)

chinchilla’s wild implications

nostalgebraist31 Jul 2022 1:18 UTC

420 points

128 comments10 min readLW link 1 review

wrapper-minds are the enemy

nostalgebraist17 Jun 2022 1:58 UTC

102 points

43 comments8 min readLW link

[Question] why assume AGIs will optimize for fixed goals?

nostalgebraist10 Jun 2022 1:28 UTC

138 points

56 comments4 min readLW link 2 reviews

autonomy: the missing AGI ingredient?

nostalgebraist25 May 2022 0:33 UTC

64 points

12 comments6 min readLW link

dalle2 comments

nostalgebraist26 Apr 2022 5:30 UTC

183 points

14 comments13 min readLW link

(nostalgebraist.tumblr.com)

[Link] Training Compute-Optimal Large Language Models

nostalgebraist31 Mar 2022 18:01 UTC

51 points

23 comments1 min readLW link

(arxiv.org)

larger language models may disappoint you [or, an eternally unfinished draft]

nostalgebraist26 Nov 2021 23:08 UTC

260 points

31 comments31 min readLW link 2 reviews

the scaling “inconsistency”: openAI’s new insight

nostalgebraist7 Nov 2020 7:40 UTC

148 points

14 comments9 min readLW link

(nostalgebraist.tumblr.com)

on “learning to summarize”

nostalgebraist12 Sep 2020 3:20 UTC

25 points

13 comments8 min readLW link

(nostalgebraist.tumblr.com)

interpreting GPT: the logit lens

nostalgebraist31 Aug 2020 2:47 UTC

223 points

37 comments11 min readLW link

is gpt-3 few-shot ready for real applications?

nostalgebraist3 Aug 2020 19:50 UTC

31 points

5 comments9 min readLW link

(nostalgebraist.tumblr.com)

[updated] how does gpt2′s training corpus capture internet discussion? not well

nostalgebraist27 Jul 2020 22:30 UTC

25 points

3 comments2 min readLW link

(nostalgebraist.tumblr.com)

[Question] Why is pseudo-alignment “worse” than other ways ML can fail to generalize?

nostalgebraist18 Jul 2020 22:54 UTC

45 points

9 comments2 min readLW link

GPT-3: a disappointing paper

nostalgebraist29 May 2020 19:06 UTC

65 points

43 comments8 min readLW link 1 review

covid-19 notes, 4/19/20

nostalgebraist20 Apr 2020 5:30 UTC

29 points

1 comment9 min readLW link

(nostalgebraist.tumblr.com)