Pure concatenation of LLM features degrades GNN accuracy on homophilous datasets, with Delta_sig metric predicting when the drop occurs better than homophily.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 2roles
background 1polarities
background 1representative citing papers
This perspective article develops a definition of foundational MLIPs and poses six open questions that the authors believe will define future research in machine-learned interatomic potentials.
citing papers explorer
-
LLM Features Can Hurt GNNs: Concatenation Interference on Homophilous Graph Benchmarks
Pure concatenation of LLM features degrades GNN accuracy on homophilous datasets, with Delta_sig metric predicting when the drop occurs better than homophily.
-
Six Open Questions in Machine-Learned Interatomic Potential Foundation Models
This perspective article develops a definition of foundational MLIPs and poses six open questions that the authors believe will define future research in machine-learned interatomic potentials.