Post Hoc Model Interpretability

Use when after training a GNN model on molecular structures with continuous targets (e.g., CCS values), when you need to understand which node-level (atom) or edge-level (bond) features contribute most to individual or aggregate predictions.

HolobiomicsLab Updated 16 repo stars
ASB pipeline: research papers, ASB, a domain benchmark of skills, tools and data, agents evaluated on it, and community validation, with a feedback loop defined

File contents

HolobiomicsLab/asb-skill-collections commit 4f9808db20

Frequently asked questions

npx get-skill add holobiomicslab/post-hoc-model-interpretability-3