Viewing profile — jncraton

You are right that the concept of "safe" is nebulous, but the goal here is specifically to be XSS-safe [1]. Elements or properties that could allow scripts to execute are removed. …

2 and 3 bit is where quality typically starts to really drop off. MXFP4 or another 4-bit quantization is often the sweet spot.

That's great. Here's "me" implementing a JS version of that library in one shot using Github Copilot and a 1 sentence prompt: > Implement when.js as a simple, zero-dependency js li…

I've adjusted or removed those sentences in the article.

Pyodide supports numpy and scipy. https://pyodide.org/en/stable/usage/packages-in-pyodide.html

The languagemodels[1] package that I maintain might meet your needs. My primary use case is education, as myself and others use this for short student projects[2] related to LLMs, …

It would be nice to see the Phind Instant weights released under a permissive license. It looks like it could be a useful tool in the local-only code model toolbox.

The speedup would not be that high in practice for folks already using speculative decoding[1]. ANPD is similar but uses a simpler and faster drafting approach. These two enhanceme…

You might be interested in "Text Embeddings Reveal (Almost) As Much As Text": > We train our model to decode text embeddings from two state-of-the-art embedding models, and also sh…

Google released the T5 paper about 5 years ago: https://arxiv.org/abs/1910.10683 This included full model weights along with a detailed description of the dataset, training process…

> PNG uses deflate. General byte-level patterns. It does not do bespoke image-specific stuff. That's not quite the whole story. PNG does include simple filters to represent a line …

There is a large amount of theoretical research on the subject of energy limits in computing. For example, Landauer's principle states any irreversible change in information requir…

You might be interested in OpenWorm: https://openworm.org/ This paper might be helpful for understanding the nervous system in particular: https://royalsocietypublishing.org/doi/10…

This is great to see. It looks like the size of the embedding vector is half the size of text-embedding-ada-002 (768 vs 1536) while providing competitive performance. This will sav…

You might be interested in TinyStories: https://arxiv.org/abs/2305.07759 > In this work, we introduce TinyStories, a synthetic dataset of short stories that only contain words that…