FinReflectKG -- EvalBench: Benchmarking Financial KG with Multi-Dimensional Evaluation

ArXi:2510.05710v2 Announce Type: replace-cross Large language models (LLMs) are increasingly being used to extract structured knowledge from unstructured financial text. Although prior studies have explored various extraction methods, there is no universal benchmark or unified evaluation framework for the construction of financial knowledge graphs (KG). We