正在加载视频...

视频加载失败

jev _can_ generate text

78,085 次观看 • 1 个月前 •via X (Twitter)

36 条评论

typedfemale 的头像
typedfemale1 个月前

some of these are art

Haoxing Du 的头像
Haoxing Du1 个月前

aww feels like GPT-2 small is back

Wocean Man 的头像
Wocean Man1 个月前

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.

RSC ☀️🌲 的头像
RSC ☀️🌲1 个月前

Building LLMs from first principles

Dennis Aumiller 的头像
Dennis Aumiller1 个月前

We're rediscovering 2019 BERT papers from first principles:

andrew blinn 的头像
andrew blinn1 个月前

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

tender 的头像
tender1 个月前

once there was an ugly barnacle ahh

Eumemic 的头像
Eumemic1 个月前

Are you giving it the entire dictionary as options?

Eumemic 的头像
Eumemic1 个月前

clever

typedfemale 的头像
typedfemale1 个月前

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

Eumemic 的头像
Eumemic1 个月前

Yeah I tried ASCII and it was horrendous lol

Incorrectish 的头像
Incorrectish1 个月前

LLMs from first principles?

Strife 的头像
Strife1 个月前

lmao

AI Psychosis 的头像
AI Psychosis1 个月前

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

Ambika Sukla 的头像
Ambika Sukla1 个月前

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.

Ege 的头像
Ege1 个月前

i hereby tap the sign

Ludvik Koutny 的头像
Ludvik Koutny1 个月前

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

haider 的头像
haider1 个月前

This is poetry

Rohan Arun 的头像
Rohan Arun1 个月前

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

emf 的头像
emf1 个月前

How?

Pillfurd 的头像
Pillfurd1 个月前

love Jev ok

🇪🇬 的头像
🇪🇬1 个月前

he speaks like trump

JustViewNews 的头像
JustViewNews1 个月前

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

Rasan-al-gaib 的头像
Rasan-al-gaib1 个月前

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

mehul 的头像
mehul1 个月前

this is giving early gpt or even gpt 2

Ahmet Karapinar 的头像
Ahmet Karapinar1 个月前

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

Izem Guendoud 的头像
Izem Guendoud1 个月前

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

Skyler Ji 的头像
Skyler Ji1 个月前

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

Coderman 🇵🇸 的头像
Coderman 🇵🇸1 个月前

Trained on Donald Trump speeches?

strider 的头像
strider1 个月前

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

Nanne 的头像
Nanne1 个月前

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

Rakib 的头像
Rakib1 个月前

what a horror

0xE  @enol.app 的头像
0xE  @enol.app1 个月前

Doesn't this show how pointless jev is?

@_@ 的头像
@_@1 个月前

hmm. needs work

Nanda 的头像
Nanda1 个月前

It's like making a car sail. Bravo 🍾

w 的头像
w1 个月前

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

相关视频

Jev has been exploding across GitHub since launch, here's what people have already built with it if you have API access and don't know where to start, copy this: jev-trader - real trading bot placing live limit orders on Monad every 300ms block, judged by Jev alone. 1,911 stars jev-ultrafast - browser agent that picks every click itself, only calling a text model when it actually needs to type something. 16,758 stars jev-doom-agent - real Chocolate Doom compiled to WebAssembly, two engines running the same map, Jev picking the tactical macro every frame jev-t-rex-runner - the Chrome dinosaur game you've procrastinated with a hundred times, now played entirely by Jev picking jump, duck, or keep running typesafe-chess - Jev vs a real search engine, two games, colors swapped. the search won both, but overruled Jev's first instinct on roughly half the moves jev-drone - a simulated quadrotor clears a five-station obstacle course by camera alone, Jev judging the situation twice a second tax-doc-classifier - sorts real IRS tax forms with 100% strict accuracy across 261 forms, at roughly $0.001 a page killmyidea - describe your startup idea, Jev scores it from every angle, then hands back kill, fix, or ship jev-curate - streams Parquet and JSONL rows through typed judgments at 1,500+ rows a second, keeping only what clears the bar pg-jev - a PostgreSQL extension that lets you ask your own database tables plain-English questions and get a real answer back none of these ten generate a single word of text. every one of them returns a number against an answer someone already defined full setup below, then run the three-question test from the article before you build an eleventh

Ryven

281,283 次观看 • 1 个月前

Jev has been exploding in popularity recently. If you already have access to the Jev API but aren't sure how to start experimenting with it, just copy this checklist: 1. jev-ultrafast Browser Use's fastest agent. Jev decides the next action and which element to click, and a language model is only called when text has to be typed. 2. typesafe-mario Jev plays Super Mario Bros. from structured emulator state, choosing every action from features pulled out of the game. 3. jev-plays-pokemon Reads Pokémon Red's game state as text, answers typed questions each turn, and lets plain code turn the answers into moves. 4. jev-drone A camera-only autonomous drone in MuJoCo with a Jev judgment model sitting in the control loop at 2.5 Hz. 5. robo-harness A real SO-101 robot arm workbench where Jev picks bounded joint steps from typed candidate actions under a spend budget. 6. fast-jev-compaction Claude Code plugin that replaces the compaction summary with Jev decisions, scoring every tool call for whether it is still needed. 7. jev-claude Routes Claude Code's own judgment calls through Jev: typed choices with probabilities at plan approval, on questions, and before risky commands. 8. is-malicious Supply-chain check before you run anything: Jev Noul checks over source and build files, returning the implicated files and lines. 9. sqlite-jev Jev inside SQL. Noul, Choice and Score judgments exposed as SQLite functions, with confidence on every row. 10. jevinci Paints images by having Jev predict every pixel's colour in parallel, with confidence deciding how wide each stroke is drawn. Copy these complete Jev blueprints - then read full Jev setup below ↓ ↓

Hanako

29,894 次观看 • 1 个月前