fi-le

6 ms·

31.
Coding Intelligence Asymptotics(fi-le.net)
20 points by fi-le 10mo ago|1 comments
32.
Where Is GPT in the Chomsky Hierarchy?(fi-le.net)
55 points by fi-le 10mo ago|76 comments
33.
World Spider Catalog(wsc.nmbe.ch)
2 points by fi-le 10mo ago|1 comments
34.
How LLMs Could Use Their Own Parameters to Hide Messages(spylab.ai)
2 points by fi-le 11mo ago|0 comments
35.
by fi-le 11mo ago
Thank you for the kind words.
36.
by fi-le 11mo ago
Thank you! As the other commenters already figured out, I manually write HTML and use tufte.css with some minor customizations.
37.
by fi-le 11mo ago
What percentage, in numbers?
38.
by fi-le 11mo ago
Evals for programming languages with formal verification. It's not clear how far we are from good coding performance in less popular languages in general, and formal verification has some quirks on top also.
39.
Operating Margins(fi-le.net)
289 points by fi-le 11mo ago|122 comments
40.
A/B Testing Could Lead LLMs to Retain Users Instead of Helping Them(newsletter.danielpaleka.com)
1 points by fi-le 11mo ago|0 comments
41.
Applied Keynesian Psychology(fi-le.net)
3 points by fi-le 11mo ago|0 comments
42.
Ten Lessons I Wish I Had Been Taught (1997) [pdf](ams.org)
2 points by fi-le 1y ago|0 comments
43.
by fi-le 1y ago
Done, thank you!
44.
by fi-le 1y ago
Hi, thanks! If someone posts better translations I will update them.
45.
What GPT-OSS leaks about OpenAI's training data(fi-le.net)
348 points by fi-le 1y ago|82 comments
46.
Hilbert's Power(fi-le.net)
2 points by fi-le 1y ago|0 comments
47.
by fi-le 1y ago
Did anyone get confirmation that the form got sent? There is no feedback from pressing "submit" for me.
48.
Trends in LLM-Generated Citations on ArXiv(spylab.ai)
1 points by fi-le 1y ago|0 comments
49.
Who Gives a Sheaf?(youtube.com)
13 points by fi-le 1y ago|0 comments
50.
That Squiggly, Treacherous Line(fi-le.net)
8 points by fi-le 1y ago|1 comments
51.
by fi-le 1y ago
Good point. The architectural solution that would come to mind is 2D text embeddings, i.e. we add 2 sines and cosines to each token embedding instead of 1. Apparently people have done it before: https://arxiv.org/abs/24
52.
by fi-le 1y ago
Two corrections, as written in the post: At least Claude not able to solve the standard levels at all, and community levels are definitely in scope.
53.
by fi-le 1y ago
At least in this instance, it came from my fleshy human brain. Although I perhaps used it to come off as smarter than I really am - just like an LLM might.
54.
Baba Is Eval(fi-le.net)
254 points by fi-le 1y ago|55 comments
55.
Handling the Shoes(fi-le.net)
2 points by fi-le 1y ago|0 comments
56.
A Database of Carbon Offset Projects(carbonplan.org)
2 points by fi-le 1y ago|0 comments
57.
by fi-le 1y ago
That is the sweetest compliment I could have hoped for, thank you.
58.
by fi-le 1y ago
I wrote it and am really confused what you mean, I think I'm missing a joke? If not this is serious, it shouldn't be there.
59.
Lessons from That 1834 Landscape Gardening Guidebook(fi-le.net)
136 points by fi-le 1y ago|26 comments
60.
Talking Well(fi-le.net)
1 points by fi-le 1y ago|0 comments

More ›