parsel VS geppetto

Parsel : A (De-)compositional Framework for Algorithmic Reasoning with Language Models

https://arxiv.org/abs/2212.10561

Here's a notebook with an introduction:

https://github.com/ezelikman/parsel/blob/main/parsel.ipynb

And here's a GUI interface the author has been developing:

http://zelikman.me/parsel/interface.html

I've been working on an augmented large language model that given these few-shot exemplars can build the below fully-functional ToDo App: ==

https://github.com/williamcotton/transynthetical-engine/tree...

https://www.williamcotton.com/articles/junie-browser-builder...

All of this is still very rough around the edges, prone to errors of various kinds, and generally not ready for prime time, but anyone is welcome to play around with what is there!

I agree. My goal, which seems very achievable from my current research, is to write formal specifications in English and generate functioning code.

One of my favorite recent projects is called Parsel:

Parsel : A (De-)compositional Framework for Algorithmic Reasoning with Language Models

https://arxiv.org/abs/2212.10561

Here's a notebook with an introduction:

https://github.com/ezelikman/parsel/blob/main/parsel.ipynb

And here's a GUI interface the author has been developing:

http://zelikman.me/parsel/interface.html

Another is, "why design a new language instead of writing this as, e.g., a Python library?" My response is we did. Internally, Parsel is in Python, and a "Function" class already exists - you can find it on GitHub. Still, you need a process to generate implementations and select one satisfying the constraints, which we call the compiler.

We made this available here!: https://github.com/ezelikman/parsel

👉 For code language models, every token is a new chance to break a program. What if LLMs wrote code like people, decomposing programs into solvable parts? They can solve competition-level coding problems by writing natural language programs in Parsel, beating prior SoTA by >75%! Quick Read: https://www.marktechpost.com/2023/01/29/researchers-at-stanford-introduce-parsel-an-artificial-intelligence-ai-framework-that-enables-automatic-implementation-and-validation-of-complex-algorithms-with-code-large-language-models-llms/ Paper: http://zelikman.me/parselpaper/parsel\_paper.pdf Github: https://github.com/ezelikman/parsel 🚀 Stay up-to-date on the latest AI news in just 3 minutes with our daily digest. Join a community of thousands of AI professionals and stay informed: https://marktechpost-newsletter.beehiiv.com/

I built the thing inside a docker container for "safety" (there is probably a lot of improvement to make on that aspect).

Here is the repo if you want to take a look: https://github.com/antca/geppetto/ It's just a WIP experiment, don't take it too seriously, please. :D