RSS

nostalgebraist

Karma: 5,666

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

nostalgebraistSep 29, 2024, 10:07 PM
246 points
41 comments11 min readLW link

in­struc­tion tun­ing and au­tore­gres­sive dis­tri­bu­tion shift

nostalgebraistSep 5, 2024, 4:53 PM
40 points
5 comments5 min readLW link

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

nostalgebraistAug 27, 2024, 11:17 PM
92 points
33 comments8 min readLW link

nos­talge­braist’s Shortform

nostalgebraistMay 27, 2024, 5:49 PM
7 points
8 comments1 min readLW link

OpenAI API base mod­els are not syco­phan­tic, at any size

nostalgebraistAug 29, 2023, 12:58 AM
183 points
20 comments2 min readLW link
(colab.research.google.com)

chin­chilla’s wild implications

nostalgebraistJul 31, 2022, 1:18 AM
424 points
128 comments10 min readLW link1 review

wrap­per-minds are the enemy

nostalgebraistJun 17, 2022, 1:58 AM
102 points
43 comments8 min readLW link

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

nostalgebraistJun 10, 2022, 1:28 AM
144 points
57 comments4 min readLW link2 reviews

au­ton­omy: the miss­ing AGI in­gre­di­ent?

nostalgebraistMay 25, 2022, 12:33 AM
69 points
12 comments6 min readLW link

dalle2 comments

nostalgebraistApr 26, 2022, 5:30 AM
183 points
14 comments13 min readLW link
(nostalgebraist.tumblr.com)

[Link] Train­ing Com­pute-Op­ti­mal Large Lan­guage Models

nostalgebraistMar 31, 2022, 6:01 PM
51 points
23 comments1 min readLW link
(arxiv.org)

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

nostalgebraistNov 26, 2021, 11:08 PM
260 points
31 comments31 min readLW link2 reviews

the scal­ing “in­con­sis­tency”: openAI’s new insight

nostalgebraistNov 7, 2020, 7:40 AM
148 points
14 comments9 min readLW link
(nostalgebraist.tumblr.com)

on “learn­ing to sum­ma­rize”

nostalgebraistSep 12, 2020, 3:20 AM
25 points
13 comments8 min readLW link
(nostalgebraist.tumblr.com)

in­ter­pret­ing GPT: the logit lens

nostalgebraistAug 31, 2020, 2:47 AM
225 points
37 comments10 min readLW link

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

nostalgebraistAug 3, 2020, 7:50 PM
31 points
5 comments9 min readLW link
(nostalgebraist.tumblr.com)

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

nostalgebraistJul 27, 2020, 10:30 PM
25 points
3 comments2 min readLW link
(nostalgebraist.tumblr.com)

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

nostalgebraistJul 18, 2020, 10:54 PM
45 points
9 comments2 min readLW link

GPT-3: a dis­ap­point­ing paper

nostalgebraistMay 29, 2020, 7:06 PM
65 points
43 comments8 min readLW link1 review

covid-19 notes, 4/​19/​20

nostalgebraistApr 20, 2020, 5:30 AM
29 points
1 comment9 min readLW link
(nostalgebraist.tumblr.com)