NumLLM: Numeric-Sensitive Large Language Model for Chinese Finance

By Huan-Yi Su, Ke Wu, Yu-Hao Huang, Wu-Jun Li

Rating

1438
Battle Count: 84

Relevance

7/10
The model's improved numeric understanding in financial contexts could be valuable for analyzing financial reports and market data

Implementation Complexity

6/10
Requires expertise in LLM fine-tuning and LoRA, but builds on existing models and techniques

Reproducibility

4/5
The paper provides detailed methodology and hyperparameters, but the full dataset is not publicly available

About this paper

Methodology: NumLLM. Problem types: Natural Language Processing, Question Answering.

The interactive Everscope explorer (charts, battles, favorites) loads below.