3345.run
running · build 0.4.2
// $ ./3345.run --start

A playground for small experiments

Animations, code snippets and interactive demos in one place. The project is starting up — something bigger will appear here soon.

Code

Tokenization

const text = "hello world";
const tokens = text
  .split(/\s+/);
// → ["hello","world"]
Python

Transformers

from transformers import pipeline
nlp = pipeline("text-generation")
nlp("Hello, ")
JS

API example

fetch('/api/llm', {
  method: 'POST',
  body: JSON.stringify({
    prompt: "Hello"
  })
})
Demo

Try the model

Model response:
Filter

Data cleanup

Stats

Parameters

context_window: 32k
params:         7.2B
latency:        ~180ms
// updated: 2026-04-23