FinAnchor: Aligned Multi-Model Representations for Financial Prediction

2026-02-24

AI Summary

Financial long-document prediction is hindered by sparse informative signals, strong noise interference, and inconsistent performance of large language models (LLMs) across tasks and time periods, which limits the effectiveness of any single model. To address this, this work proposes FinAnchor, a framework that linearly aligns and aggregates embeddings from multiple heterogeneous LLMs—without fine-tuning—into a unified representation space via anchor points. By innovatively leveraging an anchoring mechanism to fuse diverse, unmodified LLM representations, FinAnchor achieves significantly superior performance over strong single-model baselines and conventional ensemble methods across multiple financial NLP tasks, demonstrating its effectiveness in enhancing predictive generalization and stability.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Multiagent Systems: Mechanism Design

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Financial prediction from long documents involves significant challenges, as actionable signals are often sparse and obscured by noise, and the optimal LLM for generating embeddings varies across tasks and time periods. In this paper, we propose FinAnchor(Financial Anchored Representations), a lightweight framework that integrates embeddings from multiple LLMs without fine-tuning the underlying models. FinAnchor addresses the incompatibility of feature spaces by selecting an anchor embedding space and learning linear mappings to align representations from other models into this anchor. These aligned features are then aggregated to form a unified representation for downstream prediction. Across multiple financial NLP tasks, FinAnchor consistently outperforms strong single-model baselines and standard ensemble methods, demonstrating the effectiveness of anchoring heterogeneous representations for robust financial prediction.

Problem

• financial prediction
• long documents
• sparse signals
• noise
• heterogeneous LLMs

Innovation

• multi-model alignment
• financial prediction
• embedding fusion
• anchor representation
• LLM ensemble