Skip to content

Comment on QLoRA: Efficient Finetuning of Quantized LLMsparent

Comments

I'd love to hear more about your approach for getting LLMs to understand how to use plugin commands. My own experiments have not worked very well (even vanilla ChatGPT through the gpt-3.5-turbo API doesn't seem to get the concept, most of the time).

I'll do a show HN probably at the beginning of next month after this hackathon, but basically look into langchain's "tools" and the different agents they have. You don't need langchain for this at all, but it gives you the groundwork.

I saw their Code Interpreter demo on Twitter (converting an uploaded video file in a chat UI) and decided that I need that, without continuing to pay them money (because they still haven't given me access to it yet.)

So, that, and after sam a went in front of congress for the regulatory capture play, was the motivation I needed to work towards commoditizing these fuckers.

The secret sauce here with code interpreter is, well, literally a python code interpreter you can run in your browser, and it's not so secret.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.