STaR: Towards Effective and Stable Table Reasoning via Slow-Thinking Large Language Models

TL;DR

STaR introduces a slow-thinking large language model with a two-stage training process and uncertainty quantification to improve the effectiveness and stability of table reasoning, achieving state-of-the-art results.

Contribution

The paper presents a novel slow-thinking framework with a two-stage training and uncertainty measures for stable, multi-step table reasoning with LLMs.

Findings

Achieves state-of-the-art performance on in-domain benchmarks.

Demonstrates strong generalization to out-of-domain datasets.

Enhances reasoning stability through trajectory-level uncertainty quantification.

Abstract

Table reasoning with large language models (LLMs) plays a critical role in building intelligent systems capable of understanding and analyzing tabular data. Despite recent progress, existing methods still face key limitations: their reasoning processes lacks depth and explicit multi-step reasoning, often relying solely on implicit language model understanding. In addition, their reasoning processes suffer from instability, primarily caused by model uncertainty. In this work, we propose STaR, a novel slow-thinking model that can achieve effective and stable table reasoning. To enable effective multi-step reasoning, we design a two-stage training framework consisting of supervised fine-tuning (SFT) warm-up followed by reinforced fine-tuning (RFT). Specifically, in the SFT stage, we construct a high-quality dataset through automatic self-verification. In the RFT stage, we introduce a difficulty-aware reinforcement learning mechanism to further enhance reasoning capabilities. Furthermore, to improve reasoning stability, we introduce trajectory-level uncertainty quantification, which fuses token-level confidence with answer-level consistency, enabling the selection of better reasoning trajectories. Extensive experiments demonstrate that STaR-8B achieves state-of-the-art performance on in-domain benchmarks and exhibits strong generalization to out-of-domain datasets, highlighting its potential for enhancing both effectiveness and stability in table reasoning.