mo35/quant-finance-dataset

一个包含24个问答示例的数据集，专为定量金融领域的大型语言模型（LLMs）微调而设计。数据集涵盖了波动率模型（如SABR、Bergomi、rBergomi、Heston）、衍生品定价（如Dupire、VIX、Black-Scholes Greeks、CVaR）、利率与信用（如HJM、Hull-White、Merton、CDS）、数值方法（如Crank-Nicolson、Monte Carlo、FFT、LSM）以及量化策略（如Momentum、Pairs Trading、VaR、Fama-French）等多个主题。该数据集旨在改进LLMs在金融领域的推理能力，纠正定量金融模型中的常见误解，并提供紧凑、高质量的监督信号。

A dataset of 24 Q&A examples designed to fine-tune large language models (LLMs) for quantitative finance. The dataset covers various topics such as volatility models (e.g., SABR, Bergomi, rBergomi, Heston), derivatives pricing (e.g., Dupire, VIX, Black-Scholes Greeks, CVaR), interest rates & credit (e.g., HJM, Hull-White, Merton, CDS), numerical methods (e.g., Crank-Nicolson, Monte Carlo, FFT, LSM), and quant strategies (e.g., Momentum, Pairs Trading, VaR, Fama-French). It aims to improve financial reasoning in LLMs, correct common misconceptions in quantitative finance models, and provide compact, high-quality supervision signals.