FinLLM-KG Integrates Knowledge Graphs into LLMs for Financial Sentiment

FinLLM-KG is a model-agnostic framework that injects company-level knowledge from SEC 10-K filings into causal LLMs at inference time. By using natural-language profiles or factual triples, it enhances sentiment analysis without the computational cost of weight updates.