TriAgent cuts LLM costs for financial sentiment analysis

Researchers have developed TriAgent, a novel multi-agent system designed to reduce the cost of financial sentiment analysis using large language models. The system stratifies agents by contextual granularity, employing a word-level lexicon (VADER), a sentence-level transformer (FinBERT), and a cross-sentence reasoner (Qwen2.5, Mistral-7B, Phi-3.5-mini). A Semantic Divergence Index (SDI) measures disagreement between agents to route queries efficiently, achieving an F1 score of approximately 0.87 while significantly cutting costs compared to baseline LLM approaches. TriAgent also functions as a hallucination detector and enables cost-effective cross-border canonicalization of sentiment analysis. AI

IMPACT Reduces LLM operational costs for financial sentiment analysis and improves hallucination detection.

RANK_REASON Academic paper detailing a new method for LLM-based financial sentiment analysis. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• FinBERT
• GPT-4o mini
• mistral:7b
• Phi-3.5-mini
• Qwen2.5
• Semantic Divergence Index
• Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks
• Shared Consensus Dictionary
• TriAgent
• VADER

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →