E-Reverance

9 ms·

121.
Flow Map Learning via Nongradient Vector Flow [pdf](openreview.net)
26 points by E-Reverance 5mo ago|2 comments
122.
by E-Reverance 5mo ago
I think its endearing
123.
by E-Reverance 5mo ago
I might be misinterpreting but the LUAR model (which is a transformer) seems to do decently well https://www.nature.com/articles/s41599-025-06340-3/figures/2
124.
by E-Reverance 5mo ago
https://www.bbc.com/news/articles/c0mgndkklvmo
125.
by E-Reverance 5mo ago
> 75% Wouldn’t an absolute number make more sense to show than a percent. 75% is pretty good in some places in the world (not justifying the discrepancy though)
126.
by E-Reverance 5mo ago
Could this be used for batch filtering?
127.
How to Generate Text in One Step(one-step-lm.github.io)
3 points by E-Reverance 5mo ago|0 comments
128.
Neural Particle Automata: Learning Self-Organizing Particle Dynamics(arxiv.org)
1 points by E-Reverance 6mo ago|0 comments
129.
Open AI's competition spammed by AI slop(github.com)
5 points by E-Reverance 6mo ago|0 comments
130.
Effective RNG in ARC-AGI 3(xcancel.com)
2 points by E-Reverance 6mo ago|0 comments
131.
by E-Reverance 6mo ago
The first sentence basically does though, no?
132.
Matrix Valued Residuals(arxiv.org)
5 points by E-Reverance 6mo ago|0 comments
133.
GE's Walking Truck – ca. 1965(youtube.com)
2 points by E-Reverance 6mo ago|0 comments
134.
Training-Free Infinite Video Generation via Evolving Memory Tokens(memrope.github.io)
4 points by E-Reverance 6mo ago|0 comments
135.
by E-Reverance 6mo ago
I get your point but the original title kind of buries the lede (or at least the isn’t the part I personally found interesting)
136.
by E-Reverance 6mo ago
Only that high for special scenarios, but still interesting
137.
30k Tok/S (Allegedly)(percepta.ai)
11 points by E-Reverance 6mo ago|5 comments
138.
by E-Reverance 6mo ago
Happy to hear of this anecdata, as it gives me hope something similar will happen to my family
139.
Towards a Neural Debugger for Python(arxiv.org)
1 points by E-Reverance 6mo ago|0 comments
140.
by E-Reverance 6mo ago
Look very closely at the examples near the bottom, suspiciously good
141.
Rectified Flow Auto Coder(world-snapshot.github.io)
1 points by E-Reverance 6mo ago|1 comments
142.
by E-Reverance 7mo ago
SOTA right now is nano-banana 2 and luma uni-1, but a much broader point to be made is why should anyone use this service given its low quality and triviality
143.
by E-Reverance 7mo ago
Look at the text rendering in the Elizabeth one
144.
by E-Reverance 7mo ago
Not even using the SOTA models for the samples page, bit shameless no?
145.
Speculative Speculative Decoding (SSD)(arxiv.org)
61 points by E-Reverance 7mo ago|9 comments
146.
by E-Reverance 7mo ago
Not sure how much this fits into the rules but I saw on twitter someone claimed 28 params : https://gist.github.com/SeuperHakkerJa/da3050739bea97aabd86e...
147.
One-Instruction Set Computer(en.wikipedia.org)
4 points by E-Reverance 7mo ago|0 comments
148.
by E-Reverance 7mo ago
I've already immediately mentioned disambiguations in the replies to myself. The ideal response would be the llm stating the most important assumptions
149.
by E-Reverance 7mo ago
Car wash is 100 m from my house, should I drive or walk there if I need my car washed and my car is at home? Mercury 2 Today at 11:32 AM Thought for 1 seconds Driving the car is the practical choice. Since the car is at home and you need it
150.
by E-Reverance 7mo ago
"if I need my car washed" I thought adding this would push it, but I guess its prior is too strong

More ›