guardrails VS flowchart-fun

You're absolutely correct, and I agree that there's potentially a risk of quality loss. But likewise, since these are all intrinsically linked, it may be possible to leverage strength by combining these tasks. I'm unaware of a paper reviewing the reliability and/or performance of LLMs in this specific scenario. If you find any, do share :) With regards to generating JSON responses - there are simple ways to nudge the model and even validate it, using libraries such as https://github.com/promptslab/Promptify, https://github.com/eyurtsev/kor and https://github.com/ShreyaR/guardrails

There are a couple different approaches:

- Use multi-shot prompting with something like guardrails to try prompting a commercial model until it works. [1]

- Use a local model with something with a final layer that steers token selection towards syntactically valid tokens [2]

[1] https://github.com/ShreyaR/guardrails

[2] "Structural Alignment: Modifying Transformers (like GPT) to Follow a JSON Schema" @ https://github.com/newhouseb/clownfish.

👍 validate and correct the outputs of LLMs using guardrails

not vouching for it, but I know this is at least a thing that exists and I like the general idea: https://github.com/shreyar/guardrails

https://flowchart.fun recently helped a lot with a project.

- flowchart.fun - paid - https://flowchart.fun

flowchart.fun though!

On https://flowchart.fun I found that I got better overall results by asking GPT for an intermediate syntax that it was less likely to mess up (and easier for me to parse), and then parsing and transforming that syntax to my DSL. The relevant code: https://github.com/tone-row/flowchart-fun/blob/main/api/prom...

https://flowchart.fun is a good one if you're looking for something fast. (It's text-to-diagram)

That's the use-case we're angling for with our GPT integration on flowchart.fun

If you just need the image for your interface, you can generate one very quickly with https://flowchart.fun