goldemerald

6 ms·

1.
by goldemerald 29d ago
It looks like they post-trained Qwen3.6. Interesting to see how far they could improve it with they harness/algorithm.
2.
by goldemerald 3mo ago
I was able to replicate OP's attack. Since ChatGPT generates images via a separate model, I was able to ask it to tell me what the inputs to the tool was. It's a null prompt: a completely unconditional image generation. What I&#x2
3.
by goldemerald 10mo ago
Algorithmically served short form videos is clearly the smoking of our time. I cannot stand the conservative view of "well we don't know the videos cause mental health decline, or if it's simply those with a genetic inclinati
4.
Just know stuff (or, how to achieve success in a machine learning PhD) (2023)(kidger.site)
3 points by goldemerald 10mo ago|0 comments
5.
LLM-generated text is not testimony(lesswrong.com)
1 points by goldemerald 10mo ago|0 comments
6.
by goldemerald 1y ago
No discussion with Schmidhuber is complete without the infamous debate at NIPS 2016 https://youtu.be/HGYYEUSm-0Q?t=3780 . One of my goals as a ML researcher is to publish something and have Schmidhuber claim he's alrea
7.
by goldemerald 1y ago
That is addressing the incomprehensibility of PCA and applying a transformation to the entire latent space. I've never found PCA to be meaningful for deep learning. As far as I can tell, polysemous issue with neurons cannot be addresse
8.
by goldemerald 1y ago
Not quite. For an underlying semantic concept (e.g., smiling face), you can go from a basis vector [0,1,0,...,0] to the original latent space via a single rotation. You could then induce said concept by manipulating the original latent poin
9.
by goldemerald 1y ago
This is an interesting line of research but missing a key aspect: there's (almost) no references to the linear representation hypothesis. Much work on neural network interpretability lately has shown individual neurons are polysemantic
10.
by goldemerald 2y ago
I've been lightly following this type of research for a few years. I immediately recognized the broad idea as stemming from the lab of the ridiculously prolific Stefano Ermon. He's always taken a unique angle for generative models
11.
by goldemerald 2y ago
DinoV2 is an unsupervised model. It learns both a high quality global image representation and local representations with no labels. It's becoming strikingly clear that foundation models are the go to choice for common data types of na
12.
by goldemerald 2y ago
While I love XAI and am always happy to see more work in this area, I wonder if other people use the same heuristics as me when judging a random arxiv link. This paper has one author, was not written in latex, and no comment referencing a p
13.
by goldemerald 2y ago
Why not actually release the weights on huggingface? The popular SAE_lens repo has a direct way to upload the weights and there are already hundreds publicly available. The lack of training details/dataset used makes me hesitant to run
14.
Torchtitan: Large-scale LLM training using native PyTorch(github.com)
1 points by goldemerald 2y ago|0 comments
15.
by goldemerald 2y ago
"Ready to -dive- delve in?" is an amazingly hilarious reference. For those who don't know, LLMs (especially ChatGPT) use the word delve significantly more often than human created content. It's a primary tell-tale sign t
16.
Why are middlebrow dismissals so tempting? (2015)(byrnehobart.com)
1 points by goldemerald 2y ago|1 comments
17.
by goldemerald 2y ago
Solution 4 is so hilariously bad I am shocked it was suggested. Building a 2d landscape where the time dimension seems to take a random walk made laugh a lot. Ignoring the standard convention of "independent variable on x-axis" an
18.
by goldemerald 2y ago
Great work! I'd love to start using the language model variant of your work. Do you know when/if it will be open sourced? I'd start using it today if it were that soon.
19.
by goldemerald 2y ago
I like having huge wait-list so that I can put entire series on hold (like wheel of time), maybe I'm manipulating the system a little too much.
20.
Intel Gaudi 3 AI Accelerator(intel.com)
435 points by goldemerald 2y ago|250 comments
21.
by goldemerald 2y ago
I thought it was a clever/nerdy way to say in the worst case it will be out in a month. I imagine they have an internal review they have to get through first, and it's not clear if that will be done next week or in May.
22.
Drawback Chess(drawbackchess.com)
4 points by goldemerald 2y ago|0 comments
23.
by goldemerald 3y ago
Very nice. I've been working with GPT4 since it released, and I tried some of my coding tasks from today with Phind-70B. The speed, conciseness, and accuracy are very impressive. Subjectively, the answers it gives just feel better th
24.
by goldemerald 3y ago
Mostly following your advice, I got 14257. I found if I could easily eat the ghost, it was worth taking a few extra steps to go for it. The real key is knowing you can successfully leave 6 on either side to carefully pick up after getting t
25.
Ask HN: Best PDF Text Extractor?
2 points by goldemerald 3y ago|1 comments
26.
by goldemerald 3y ago
The study only compares heart attack increase over a couple days, not weeks. Once you account for longer time period after DST, there is no statistically significant difference between post DST heart attacks and the rest of the year. The ex
27.
by goldemerald 3y ago
Given all the ads builtin to the site about using generative AI, I'm inclined to think so too.
28.
How I draw figures for my mathematical lecture notes using Inkscape (2019)(castel.dev)
3 points by goldemerald 3y ago|0 comments
29.
Better Bash History (2012)(blog.sanctum.geek.nz)
3 points by goldemerald 3y ago|0 comments
30.
by goldemerald 3y ago
It's immediately obvious to me that multiple paragraphs of this article are written by an LLM. I've read (and enjoyed) this book. I especially like making chatgpt and other LLMs write about it in different perspectives. I'm s

More ›