natolambert

8 ms·

1.
Learning to solve hard problems in RL for LLMs by never giving up(mnoukhov.github.io)
52 points by natolambert 16d ago|0 comments
2.
Show HN: Colloquium – a Markdown-native slide tool for academics(github.com)
2 points by natolambert 7mo ago|0 comments
3.
The Atom Project(atomproject.ai)
13 points by natolambert 1y ago|4 comments
4.
by natolambert 1y ago
ATOM: American Truly Open Models
5.
by natolambert 2y ago
Author here! Just wanted to say that this is indeed in a good place to share, some very useful stuff, but is also very work in progress. I'm may 60% or so to my first draft. Said progress is coming every day and I happily welcome fixes
6.
by natolambert 2y ago
As the other commenter said, R1 required very standard RLHF techniques too. But a fun way to think about it is that reasoning models are going to be bigger and uplift the RLHF boat. But we need a few years to establish basics before I can w
7.
by natolambert 2y ago
thx, you're right. I've fixed it!
8.
by natolambert 3y ago
Author here, you can find more on my website: https://natolambert.com/cv Have been building RLHF systems at HuggingFace since ChatGPT, with some other experience before.
9.
How RLHF Works(interconnects.ai)
165 points by natolambert 3y ago|32 comments
10.
Different Development Paths of LLMs(interconnects.ai)
1 points by natolambert 3y ago|0 comments
11.
Evaluating and Uncovering Open LLMs(interconnects.ai)
1 points by natolambert 3y ago|0 comments
12.
Unfortunately, OpenAI and Google have moats(interconnects.ai)
2 points by natolambert 3y ago|0 comments
13.
Specifying Hallucinations of LLMs(interconnects.ai)
1 points by natolambert 3y ago|0 comments
14.
The last reliable path into AI research(natolambert.com)
1 points by natolambert 5y ago|0 comments
15.
Remote Robotic-Data Farms(robotic.substack.com)
1 points by natolambert 5y ago|0 comments
16.
Reward Is Not Enough(robotic.substack.com)
2 points by natolambert 5y ago|0 comments
17.
All machine learning becomes reinforcement learning(robotic.substack.com)
1 points by natolambert 5y ago|0 comments
18.
Debugging model-based reinforcement learning systems(natolambert.com)
2 points by natolambert 5y ago|0 comments
19.
Setting ourselves up for exploitation: RL in the wild(robotic.substack.com)
1 points by natolambert 6y ago|0 comments
20.
Decoupling AI from the latent variable of spoken languages(robotic.substack.com)
2 points by natolambert 6y ago|0 comments
21.
Covid didn’t give us personal robots, it gave us Woebot(robotic.substack.com)
1 points by natolambert 6y ago|0 comments
22.
Boston Dynamics: Studying Athletic Intelligence(robotic.substack.com)
1 points by natolambert 6y ago|0 comments
23.
Robotic Startups 2.0: Horizontal Modularity(democraticrobots.substack.com)
1 points by natolambert 6y ago|0 comments
24.
Social Networks and Degradation to the Public Square of Discourse(democraticrobots.substack.com)
2 points by natolambert 6y ago|1 comments
25.
Our idea of Free Will can bias AIs we create(democraticrobots.substack.com)
2 points by natolambert 6y ago|0 comments
26.
The Ubiquity and Future of Model-Based Reinforcement Learning(democraticrobots.substack.com)
2 points by natolambert 6y ago|0 comments
27.
Constructing Axes for (Legal) Reinforcement Learning Policy(democraticrobots.substack.com)
2 points by natolambert 6y ago|0 comments
28.
The Collingridge Dilemma and Current Policy on Robots(democraticrobots.substack.com)
2 points by natolambert 6y ago|0 comments
29.
EE's See the World in Models(democraticrobots.substack.com)
2 points by natolambert 6y ago|0 comments
30.
Automated: The levers tech companies pull to direct our lives(democraticrobots.substack.com)
1 points by natolambert 6y ago|0 comments

More ›