JT-DA: Enhancing Data Analysis with Tool-Integrated Table Reasoning Large Language Models

TL;DR

This paper introduces JT-DA-8B, a large language model specialized for complex table reasoning, trained on a diverse dataset with a multi-step reasoning workflow, achieving strong performance in real-world scenarios.

Contribution

The work presents a new specialized large language model for table reasoning, with a comprehensive training corpus and a four-stage reasoning workflow to enhance interpretability and accuracy.

Findings

Achieves strong performance on various table reasoning tasks.

Demonstrates effectiveness of data-centric generation and workflow optimization.

Utilizes a large, diverse dataset of 3 million tables for training.

Abstract

In this work, we present JT-DA-8B (JiuTian Data Analyst 8B), a specialized large language model designed for complex table reasoning tasks across diverse real-world scenarios. To address the lack of high-quality supervision in tabular reasoning scenarios, we construct a comprehensive and diverse training corpus with 34 well-defined table reasoning tasks, by aggregating 29 public table QA datasets and 3 million tables. An automatic pipeline is proposed to generate realistic multi-step analytical tasks involving reasoning patterns. The model is trained upon open-source JT-Coder-8B model, an 8B-parameter decoder-only foundation model trained from scratch. In the training stage, we leverage LLM-based scoring and workflow-aligned filtering to distill high-quality, table-centric data. Both supervised fine-tuning (SFT) and Reinforcement learning (RL) are adopted to optimize our model. Afterwards, a four-stage table reasoning workflow is proposed, including table preprocessing, table sensing, tool-integrated reasoning, and prompt engineering, to improve model interpretability and execution accuracy. Experimental results show that JT-DA-8B achieves strong performance in various table reasoning tasks, demonstrating the effectiveness of data-centric generation and workflow-driven optimization.