Plasma GraphRAG automates physics-grounded parameter selection for gyrokinetic simulations via a domain-specific knowledge graph and LLMs, reporting over 10% better quality and up to 25% fewer hallucinations than standard RAG.
5d neural surrogates for nonlinear gyrokinetic simulations of plasma turbulence
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
physics.plasm-ph 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Plasma GraphRAG: Physics-Grounded Parameter Selection for Gyrokinetic Simulations
Plasma GraphRAG automates physics-grounded parameter selection for gyrokinetic simulations via a domain-specific knowledge graph and LLMs, reporting over 10% better quality and up to 25% fewer hallucinations than standard RAG.