For now we are storing the embeddings of the generated docstrings, the intuition behind this is kinda like how HyDE works. I don't think the actual code itself is reversible from the embeddings, but yes we do want to store the actual code at some point and we will need the proper security measures for that. We are also actively developing an self-hostable version for enterprise.

> embedding vectors you've calculated from the code? If so, those are likely quite easily reversible

I don't think embeddings are generally reversible... you're usually projecting onto a lower dimensional space, and therefore losing information.

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 show that our model can recover important personal information (full names) from a dataset of clinical notes.

https://arxiv.org/pdf/2310.06816.pdf

There's certainly information loss, but there is also a lot of information still present.

Yeah, that paper is what I was thinking about. https://simonwillison.net/2024/Jan/8/text-embeddings-reveal-...

“a multi-step method that iteratively corrects and re-embeds text is able to recover 92% of 32-token text inputs exactly”.

"Quite easily" isn't true in most cases, but embeddings are sometimes reversible. We know this because programs like Stable Diffusion sometimes output near-perfect copies of training data when given the correct prompt, and generation of that image is based on word and image embeddings alone.

I’ve never heard of reversible embeddings in practice.

In theory if you know the model being used you could reverse them.