Jev-Leftpad
github.com
github.com
important to note that the "confidence" score is... maybe not what people think it is - kind of useless, and just a convenience step from the probabilities.
from the docs: "confidence is a statistic computed from the probability distribution the answer already gives you." [0] I actually encourage people to visit the docs because it has a specific page on this with a little applet to really make this clear.
What is the use case for a classifier that works 90% of the time...? I feel like if I'm classifying something, I probably care enough that 90% ain't gonna cut it...
I guess the answer is just agential stuff that effectively gets double checked by the LLM in the driver seat, anyway? That tracks, though it means that jev is mostly just for the people making harnesses. Which is all of us but still!
Which begs the question, can Jev retest until it gets the right result? Can it tell how corellated two of its results are? 90% correct makes for a wonderful iterator, but a poor oracle.
I guess, in the end: I think it'll end up being fantastically useful for artificial engineers with their vastly superior ability to keep track of fine details and rapidly context switch, but fairly niche for any of us organic engineers that are left.
All that doesn't apply to low stakes stuff like games, though -- can't wait for the first truly open world game, NGL. A silver lining to the cobalt cloud?
"""
def fizz_buzz_encode(i):
if i % 15 == 0: return np.array([0, 0, 0, 1])
elif i % 5 == 0: return np.array([0, 0, 1, 0])
elif i % 3 == 0: return np.array([0, 1, 0, 0])
else: return np.array([1, 0, 0, 0])
me: That's enough setup, you're exactly right. [<--- !!] [...]
"""
Damn, Claude was there all along
So that's where in the training set current models get that phrase! /j
It works great, but maybe your implementation could save me some money. I’ll test it and report back.
The tests mock Jev.
I'll raise a PR which uses Jev to check if the target length is beyond this range
You pad the text.
Oh, my god.
Tech doesn't teach it's own history in a useful way, so we keep repeating it too.
it('trusts Jev when it chooses the wrong amount', async () => {
answer('space_2');
await expect(leftPad('jev', 8)).resolves.toBe(' jev');
});5 to 10 years from now, after this has worked itself deep into the npm dependency chain, we'll be lamenting how Jev-Leftpad is causing outages in critical services.
so probably early bird
Would have been funnier if it used the GLWTPL: https://spdx.org/licenses/GLWTPL.html
jev should only have the choice of space_0 or space_1, then recurse on n-1
this extends the implementation to infinite padding and is cleaner code
I feel for the people whose 'AI strategy' is keeping up with every vibecoded vibecoding junk that releases
Btw, this model also has very tiny inherent bias: https://jev-bias-analysis.stupidlabs.lol/
I use it mainly to check whether this code fits the rules I defined, just a yes or no. But I'm not sure if that's the right way to use it.
And yet we learned nothing from it and all the same problems keep popping up, only this time in the form of supply chain attacks.
Leftpad is a npm package that chooses if it should or shouldn't pad the left side of a string. It was famous for bringing down everyone's npm installs a few years ago.
This combination is a double joke. Put something stupid in something stupid.
> Jev is a new type of model that just makes decisions based on given options.
Though, again, a yet another project for a yet another "AI" to make someone else more dependent on it...
Related:
https://en.wikipedia.org/wiki/Jev_(AI_model)
https://typesafe.ai/