Rendered at 17:50:25 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
hhjinks 11 hours ago [-]
Does the mapping depend on the chose set, or could you just map all 25k words and pick 10 points that form a circle? From the process, they seemingly replaced single words in the set that didn't fit, which makes it sound like the position of the remaining words didn't change much from changing the set.
jamie-simon 5 hours ago [-]
interesting. yeah, there's probs some optimization that can be done there.
ZeroDayDreamer 9 hours ago [-]
Using high-dimensional statistics for beer money, not academic papers" is the correct use of the degree. The papers are where you prove the circle exists. The bar is where you find out it exists for everything
mph1027 12 hours ago [-]
The best bar bets are the ones where you're technically not cheating, just abusing linear algebra.
mdritch 16 hours ago [-]
I recently did some analysis on how different style prompts like "be succinct" and "avoid mannered prose" affect LLM outputs and I also saw a ~circular relationship with a 2d dimensional reduction (MDS) of the similarity matrix. I wonder if it's some artifact of PCA/MDS + Gemma's output dist
ViscountPenguin 16 hours ago [-]
A fun step-up would be to see if you can find an infinity shape by repeating the process with a set of points which are maximally far away from all points but 1 on the existing circle.
ovin_dal 13 hours ago [-]
Betting on this with high-dimensional stats sounds like a classic case of overfitting your opponent. Just buy them a beer.
heronbank 14 hours ago [-]
My stats professor would be proud I'm using high-dimensional statistics for beer money, not academic papers.
cold_boot 14 hours ago [-]
Tried something similar predicting poker outcomes, ended up buying everyone a round. The math wasn't in my favor.
Eastmill 14 hours ago [-]
Used a similar trick once to predict who'd buy coffee. Didn't win a beer, but definitely earned some weird looks.