Unlocking Data Value in Finance: A Study on Distillation and Difficulty-Aware Training

TL;DR

This paper presents a comprehensive study on enhancing financial language models through multi-stage distillation and difficulty-aware training, resulting in superior performance on diverse financial tasks.

Contribution

Introduces high-quality, difficulty-aware datasets and training methods that improve the robustness and accuracy of financial language models.

Findings

High-quality chain-of-thought distillation improves foundation robustness.

Difficulty-aware sampling enhances RL generalization.

Models outperform state-of-the-art financial LLMs of similar size.

Abstract

Large Language Models (LLMs) have demonstrated strong general capabilities, yet their deployment in finance remains challenging due to dense domain-specific terminology, stringent numerical reasoning requirements, and low tolerance for factual errors. We conduct a controlled empirical study showing that in specialized vertical domains, performance is largely determined by the quality and difficulty/verifiability profile of post-training data. We introduce \textbf{ODA-Fin-SFT-318k}, constructed via multi-stage distillation and verification to produce high-quality Chain-of-Thought supervision, and \textbf{ODA-Fin-RL-12k}, curated for hard-but-verifiable tasks that balance reward precision and task diversity. Using standard SFT and RL pipelines, we show that high-quality CoT distillation establishes a robust foundation during SFT, while difficulty- and verifiability-aware sampling improves RL generalization. Evaluated on nine benchmarks spanning general financial tasks, sentiment analysis, and numerical reasoning, our ODA-Fin-RL-8B consistently surpasses open-source state-of-the-art (SOTA) financial LLMs of comparable size. We release our ODA-Fin-SFT-318k and ODA-Fin-RL-12k datasets, along with trained models to advance data-centric financial AI research.