Efficient Table QA via TableGrid Navigation and Progressive Inference Prompting
About
Large Language Models (LLMs) have shown promising results on NLP tasks, however, their performance on tabular data still needs research attention, because Table Question-Answering (TQA) requires precise cell retrieval and multi-step structured reasoning. Existing work improves TQA either by fine-tuning or training LLMs on task-specific tabular data, but often lacks verifiable control over how the model navigates tables and derives answers. In this work, we propose a training-free TQA approach with two structured prompting frameworks: TableGrid Navigation (TGN), which iteratively navigates rows and columns via a three-module loop to locate evidence and refine answers, and Progressive Inference Prompting (PIP), which enforces columns identification for explicit progressive row selection constraint according to the query. We evaluate 17 LLMs against 6 baselines on TableBench and FeTaQa dataset. On TableBench, TGN improves over the strongest baseline by 3.8 points, and on FeTaQa, PIP achieves SOTA performance over ReAct and Chain-of-Thought. Beyond inference-time gains, PIP and TGN can also serve as supervision templates to fine-tune small models, narrowing the performance gap to much larger architectures in resource-constrained settings, offering versatile and cost-efficient solution for TQA.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Fact Checking | TableBench (test) | Accuracy85.42 | 136 | |
| Table Question Answering | TableBench 1.0 (test) | Accuracy48.46 | 136 | |
| Numerical Reasoning | TableBench (test) | Accuracy64.48 | 13 | |
| Data Analysis | TableBench (test) | Accuracy26.63 | 10 | |
| Table Question Answering | TableBench (test) | Overall Accuracy0.4846 | 10 | |
| Table Question Answering | FeTaQA (test) | sacreBLEU19.32 | 8 |