Archive
Sequences
About
Search
Log In
Questions
Events
Shortform
Alignment Forum
AF Comments
Home
Featured
All
Tags
Recent
Comments
RSS
Robert_AIZI
Karma:
1,143
All
Posts
Comments
New
Top
Old
Page
1
Comparing Anthropic’s Dictionary Learning to Ours
Robert_AIZI
7 Oct 2023 23:30 UTC
136
points
8
comments
4
min read
LW
link
I was Wrong, Simulator Theory is Real
Robert_AIZI
26 Apr 2023 17:45 UTC
75
points
7
comments
3
min read
LW
link
(aizi.substack.com)
The Toxoplasma of AGI Doom and Capabilities?
Robert_AIZI
24 Apr 2023 18:11 UTC
68
points
12
comments
1
min read
LW
link
Why do we assume there is a “real” shoggoth behind the LLM? Why not masks all the way down?
Robert_AIZI
9 Mar 2023 17:28 UTC
61
points
48
comments
2
min read
LW
link
Research Report: Sparse Autoencoders find only 9/180 board state features in OthelloGPT
Robert_AIZI
5 Mar 2024 13:55 UTC
53
points
24
comments
10
min read
LW
link
(aizi.substack.com)
My Reservations about Discovering Latent Knowledge (Burns, Ye, et al)
Robert_AIZI
27 Dec 2022 17:27 UTC
50
points
0
comments
4
min read
LW
link
(aizi.substack.com)
GPT-4: What we (I) know about it
Robert_AIZI
15 Mar 2023 20:12 UTC
40
points
29
comments
12
min read
LW
link
(aizi.substack.com)
How does GPT-3 spend its 175B parameters?
Robert_AIZI
13 Jan 2023 19:21 UTC
40
points
13
comments
6
min read
LW
link
(aizi.substack.com)
No Really, Attention is ALL You Need—Attention can do feedforward networks
Robert_AIZI
31 Jan 2023 18:48 UTC
29
points
7
comments
6
min read
LW
link
(aizi.substack.com)
Invocations: The Other Capabilities Overhang?
Robert_AIZI
4 Apr 2023 13:38 UTC
29
points
4
comments
4
min read
LW
link
(aizi.substack.com)
Is behavioral safety “solved” in non-adversarial conditions?
Robert_AIZI
25 May 2023 17:56 UTC
26
points
8
comments
2
min read
LW
link
(aizi.substack.com)
[Question]
Question for Prediction Market people: where is the money supposed to come from?
Robert_AIZI
8 Jun 2023 13:58 UTC
25
points
26
comments
1
min read
LW
link
[Research Update] Sparse Autoencoder features are bimodal
Robert_AIZI
22 Jun 2023 13:15 UTC
23
points
1
comment
5
min read
LW
link
(aizi.substack.com)
Log-odds are better than Probabilities
Robert_AIZI
12 Dec 2022 20:10 UTC
22
points
4
comments
4
min read
LW
link
(aizi.substack.com)
Rating my AI Predictions
Robert_AIZI
21 Dec 2023 14:07 UTC
22
points
5
comments
2
min read
LW
link
(aizi.substack.com)
Research Report: Incorrectness Cascades
Robert_AIZI
14 Apr 2023 12:49 UTC
19
points
0
comments
10
min read
LW
link
(aizi.substack.com)
Early Results: Do LLMs complete false equations with false equations?
Robert_AIZI
30 Mar 2023 20:14 UTC
14
points
0
comments
4
min read
LW
link
(aizi.substack.com)
Article Review: Discovering Latent Knowledge (Burns, Ye, et al)
Robert_AIZI
22 Dec 2022 18:16 UTC
13
points
4
comments
6
min read
LW
link
(aizi.substack.com)
Unsafe AI as Dynamical Systems
Robert_AIZI
14 Jul 2023 15:31 UTC
11
points
0
comments
3
min read
LW
link
(aizi.substack.com)
Addendum: More Efficient FFNs via Attention
Robert_AIZI
6 Feb 2023 18:55 UTC
10
points
2
comments
5
min read
LW
link
(aizi.substack.com)
Back to top
Next