H2GLM combines LLM-inferred function roles with hierarchical heterogeneous GNN message passing to detect and localize malicious Python packages more accurately than prior ML, graph, and LLM baselines.
Pvdetector: Pretrained vulnerability detection on vulnerability-enriched code semantic graph,
1 Pith paper cite this work, alongside 1 external citations. Polarity classification is still indexing.
1
Pith paper citing it
1
external citations · OpenAlex
fields
cs.CR 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
LLM-Enhanced Hierarchical Heterogeneous Graph Representation Learning for Malicious Python Package Detection
H2GLM combines LLM-inferred function roles with hierarchical heterogeneous GNN message passing to detect and localize malicious Python packages more accurately than prior ML, graph, and LLM baselines.