turtlesoup

6 ms·

1.
Robloquake: Porting the Quake Engine to Roblox(github.com)
8 points by turtlesoup 2mo ago|0 comments
2.
by turtlesoup 3mo ago
Thanks! It is part of the clustering step, I tell the model to make a judgement of whether something is inanimate or hallucinated (as defined by low support from only non-frontier models / judgement). I iterated on this a lot and made
3.
by turtlesoup 3mo ago
Top percentile of people searched!
4.
by turtlesoup 3mo ago
We were drawing inspiration from https://www.michaelfogleman.com/static/nes/
5.
by turtlesoup 3mo ago
Whoops we had some scrolling bugs with sound, hopefully fixed
6.
by turtlesoup 3mo ago
Interesting, that probably reveals something about the training set for most of these models
7.
by turtlesoup 3mo ago
What tracking cookies are you seeing? The intention was just some cloudflare checks for spam identification
8.
by turtlesoup 3mo ago
This is directional; models self-report confidence on their answers and the strength is a linear combination of the confidence plus a bonus for every model that got clustered in. Models are notoriously uncalibrated especially for self-repor
9.
by turtlesoup 3mo ago
Prompt for rollouts posted below ( https://news.ycombinator.com/item?id=48592415 ). I have a bit more information on the clustering part in https://intheweights.com/about but every thing returned by the model
10.
by turtlesoup 3mo ago
Sure thing! It is the same prompt for every model in the rollouts, here it is No tools are available. Do not imply that you searched, looked up, browsed, or verified anything externally. If the name is ambiguous, return distinct likely
11.
by turtlesoup 3mo ago
No purpose, just a fun hack and science experiment. Glad to see it getting a good reception!
12.
by turtlesoup 3mo ago
Just disabled latest!
13.
by turtlesoup 3mo ago
Just deployed a fix for this; removed latest and capped pagination.
14.
by turtlesoup 3mo ago
Ah yeah, the "hallucinations" classification is optimized for recall (keeping as many results as I can) not precision. It is mostly based on small models being the only support for a claim. Certainly lots of hallucinations everywh
15.
by turtlesoup 3mo ago
Not cheap for sure but it's all for fun! I have done some optimizations to try to get cost as low as possible; the final clustering actually uses Kimi K2 for this reason. More info on https://intheweights.com/about
16.
by turtlesoup 3mo ago
It is on a 10 minute interval and only does images for the top people, should pick up Elvis shortly. On refusal it shows an X for the person, sometimes the upstream model (gpt-5.4 image 2) will refuse and there are a few names I manually om
17.
by turtlesoup 3mo ago
Didn't expect to hit the front page! Trying my best to keep it up
18.
by turtlesoup 3mo ago
Yep, those are from "the weights" of GPT-5.4 Image 2 with a little "draw <name>" query and a style reference. More details here https://intheweights.com/about
19.
by turtlesoup 3mo ago
Aye... right now the clusterer does the classification of whether it thinks it is a hallucination or not (it is biased against only small model support) but I tried to optimize for recall over precision. The query is essentially "Who i
20.
by turtlesoup 3mo ago
We need a name for these pure hallucinations, something like lucies or looseys Usually the hallucinations have some logic to them like a person with a similar spelling in some of the training sets. LLMs are mysterious!
21.
Show HN: Are You in the Weights?(intheweights.com)
471 points by turtlesoup 3mo ago|247 comments
22.
Existing in an un-simulated reality(iahwrites.substack.com)
52 points by turtlesoup 3y ago|83 comments
23.
by turtlesoup 3y ago
Author here; it's nice to see people using TWDNE even in the era of Midjourney v5 and GPT4. Always happy to answer questions if people have any!
24.
by turtlesoup 5y ago
Yup, it checks against as many dictionaries as I could find but it is non-exhaustive since language is so... flexible.
25.
by turtlesoup 5y ago
I call it the "too lazy to set up a database" pattern (I'm the original author :P). There should be an signature in there too so hopefully tamper-proof.
26.
by turtlesoup 5y ago
I haven't either (beyond anime / animals, which is widely explored ala https://www.thiswaifudoesnotexist.net/ ) -- if you do see anything let me know. Tangentially related, one more of my favourite parts of this wo
27.
by turtlesoup 5y ago
I'd suggest this code to follow: https://github.com/lucidrains/stylegan2-pytorch/blob/fc22408... . The standard practice is to interpolate between the source latents and target latents and feed the interp
28.
by turtlesoup 5y ago
Both definitions and words are machine generated!
29.
by turtlesoup 5y ago
Ah crap, good call. I did a first-pass at filtering for the ones displayed on the site but my regex must have had some misses.
30.
by turtlesoup 5y ago
Press "write your own" and it'll make up a definition for you :)

More ›