Video wird geladen...
Video konnte nicht geladen werden
jev _can_ generate text
78,085 Aufrufe • vor 1 Monat •via X (Twitter)
36 Kommentare

some of these are art

aww feels like GPT-2 small is back

omg you should cut text into small packets called tokens, and have a big array where every index represents a possible token, and then jev takes some text in and ranks the most likely next tokens in a token stream. could be revolutionary.

Building LLMs from first principles

We're rediscovering 2019 BERT papers from first principles:

this is not real text. only language models have the intentionality to generate real text. these are just sparkling stochastic enums

once there was an ugly barnacle ahh

Are you giving it the entire dictionary as options?

clever

i had claude try a few different methods (character does work, nor does token because the 256 limit)

Yeah I tried ASCII and it was horrendous lol

LLMs from first principles?

lmao

The only way I could do that was if you had to do a lot more work and then you would be done by yourself I would say I could get Twitter

Really surprised with the response about Jev which simply brings back how lms were built and trained with tasks - the very days. It really show how little everyone knows about the history of lms.

i hereby tap the sign

It sounds like Trump's brain is running on Jev :D

This is poetry

Cool I created a similar autocomplete the actually accelerates writing based on your past 1000 sentences

How?

love Jev ok

he speaks like trump

Only 255 words per request 😂. It is caveman speech.

people rediscovering that there can be usecases for transformers other than llms will always be funny to me.

this is giving early gpt or even gpt 2

@zacharynado tried character doesn’t work and with words you are only limited to the typesafeai’s max choice which is 255

pretty slow but it’s excellent as a reasoning layer for llm’s.

Pressing the middle option for autocomplete on your keyboard will do the same thing

Trained on Donald Trump speeches?

this jev is from derek zoolander center for llms who can't read good

1. Jev doesn't support input caching yet. 2. Better to give Jev a 'llm_text' option for switching to real LLM.

what a horror

Doesn't this show how pointless jev is?

hmm. needs work

It's like making a car sail. Bravo 🍾

that output text is giving me some dejavu.......
