3 ms·
Agreed 100%. DSPy along with libraries inspired by it (i.e. https://github.com/zou-group/textgrad https://github.com/zou-group/textgrad) are nothing more than f
by Der_Einzige 2y ago
Agreed 100%. DSPy along with libraries inspired by it (i.e. https://github.com/zou-group/textgrad https://github.com/zou-group/textgrad) are nothing more than fancy prompt chains under the hood.
These libraries mostly exist as "cope" for the fact that we don't have good fine-tuning (i.e. lora) capabilities for ChatGPT et al, so we try to instead optimize the prompt.
- qeternity 2y agoGlad to see others saying this. I haven't looked at it in some months, but I previously realized it's mostly a very complicated way to optimize few-shot learning prompts. It's hardly whatever magical blackbox optimizer they try to market it as.
- dmarchand90 2y agoMy guess is it will be like pascal or smalltalk, an important development for illustrating a concept but is ultimately replaced by something more rigorous
- isaacbmiller 2y ago> These libraries mostly exist as "cope" > nothing more than fancy prompt chains under the hood Some approaches using steering vectors, clever ways of fine-tuning, transfer decoding, some tree search sampling-esque approaches, and others all seem very promising. DSPy is, yes, ultimately a fancy prompt chain. Even once we integrate some of the other approaches, I don't think it becomes a single-lever problem where we can only change one thing(e.g., fine-tune a model) and that solves all of our problems. It will likely always be a combination of the few most powerful levers to pull.
- Der_Einzige 2y agoCorrect, when I say "ChatGPT et al", I mean closed source paywalled LLMs, open access LLM personalization is an extreme gamechanger. All of what you mentioned is important, and I'm particularly excited about PyReft. https://github.com/stanfordnlp/pyreft https://github.com/stanfordnlp/pyreft Anything Christopher Manning touches turns to gold.