Sparks of Tabular Reasoning via Text2SQL Reinforcement Learning

Josefa Lia Stoisser, Marc Boubnovski Martell, Julien Fauqueur

Keywords: Tabular reasoning, GRPO, text2sql

TL;DR: This study enhances Text-to-SQL tasks by teaching LLMs to reason over tabular data using chain-of-thought traces and Group Relative Policy Optimization, achieving 33.9% accuracy gain for LLaMA and 14.5% for Qwen.

Abstract:

This work reframes the Text-to-SQL task as a pathway for teaching large language models (LLMs) to reason over and manipulate tabular data—moving beyond the traditional focus on query generation. We propose a two-stage framework that leverages SQL supervision to develop transferable table reasoning capabilities. First, we synthesize detailed chain-of-thought (CoT) traces from real-world SQL queries, providing step-by-step, clause-level supervision that teaches the model how to traverse, filter, and aggregate table fields. Second, we introduce a Group Relative Policy Optimization (GRPO) reinforcement learning objective that connects SQL execution accuracy to generalizable reasoning by encouraging steps that extend beyond task-specific syntax and transfer across datasets.

Empirically, our approach improves performance on standard Text-to-SQL benchmarks and achieves substantial gains on reasoning-intensive datasets such as BIRD, CRT-QA and Tablebench, demonstrating enhanced generalization and interpretability. Specifically, the distilled-quantized LLaMA-8B model achieved a 34\% relative increase in exact match scores on CRT-QA when trained on Text-to-SQL tasks, while Qwen-2.5-7B achieved a 10\% and Qwen-2.5-14B a 6\% relative increase. These results suggest that SQL can serve not only as a target formalism but also as an effective scaffold for learning robust, transferable reasoning over structured data.

Include In Proceedings: Yes

Submission Number: 29