Suggestion: WFGY as a robustness / debugging resource for LLM and RAG systems #4
Description
Hi, thanks for maintaining this collection of ML software, tutorials and datasets.
I would like to suggest WFGY as a resource in the LLM or tools section.
- Project: WFGY
- Repo: https://github.com/onestardao/WFGY
- License: MIT
- GitHub: 1.4k+ stars
Short description:
- WFGY is a framework for debugging and stabilizing LLM based systems, especially RAG pipelines.
- It is used together with existing libraries and vector stores, not instead of them.
Key resource:
ProblemMap (16 failure modes)
https://github.com/onestardao/WFGY/tree/main/ProblemMap
- Each failure mode has a dedicated page with explanation, typical logs or symptoms, and suggestions on how to repair the system.
- The goal is to promote systematic debugging and to avoid “just tweak the prompt again”.
There is also an advanced part:
- WFGY 3.0 Singularity Demo, a TXT based evaluation pack with 131 high difficulty questions for LLMs.
If this fits your collection, I can send a PR with a short entry.