Research map / Large language models / Agents, code and math
SheetBrain: A Neuro-Symbolic Agent for Accurate Reasoning over Complex and Large Spreadsheets
Ziwei Wang, Jiayuan Su, Mengyu Zhou, Huaxing Zeng, Mengni Jia, Xiao Lv, Haoyu Dong, Xiaojun Ma, Shi Han, Dongmei Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add understanding module (ablation) | SheetBrain · RealHiTBench | hit rate (%): 75.0 -> 78.3 | Improved |
| add understanding module (ablation) | SheetBrain · SheetBench | hit rate (%): 77.0 -> 80.3 | Improved |
| add validation module (ablation) | SheetBrain · RealHiTBench | hit rate (%): 76.7 -> 78.3 | Improved |
| add validation module (ablation) | SheetBrain · SheetBench | hit rate (%): 77.0 -> 80.3 | Improved |
| add understanding and validation modules together (ablation) | SheetBrain · SheetBench | hit rate (%): 73.8 -> 80.3 | Improved |
| add inspector tool (ablation) | SheetBrain execution module · SheetBench | accuracy (%): 77.3 -> 79.1 | Improved |
| add search tool (ablation) | SheetBrain execution module · SheetBench | accuracy (%): 73.1 -> 79.1 | Improved |
| use code sandbox instead of JSON-based neural calling (ablation) | SheetBrain execution module · SheetBench | accuracy (%): 65.1 -> 79.1 | Improved |
| add cell position information to Markdown serialization (ablation) | SheetBrain · table encoding | accuracy: 63.3 -> 75.0 | Improved |
| add cell position information to HTML serialization (ablation) | SheetBrain · table encoding | accuracy: 59.7 -> 76.3 | Improved |
| use SheetBrain instead of vanilla GPT-4.1 | GPT-4.1 · MultiHiertt | accuracy (%): 53.5 -> 62.6 | Improved |
| use SheetBrain instead of vanilla GPT-4.1 | GPT-4.1 · RealHiTBench | accuracy (%): 70.0 -> 78.3 | Improved |
Closest papers by meaning
- SpreadsheetBench: Towards Challenging Real World Spreadsheet Manipulation (NeurIPS 2024 · similarity 0.87)
- NeedleInATable: Exploring Long-Context Capability of Large Language Models towards Long-Structured Tables (NeurIPS 2025 · similarity 0.86)
- MMTU: A Massive Multi-Task Table Understanding and Reasoning Benchmark (NeurIPS 2025 · similarity 0.86)
- CodePlan: Unlocking Reasoning Potential in Large Language Models by Scaling Code-form Planning (ICLR 2025 · similarity 0.85)
- LogicCat: A Chain-of-Thought Text-to-SQL Benchmark for Complex Reasoning (AAAI 2026 · similarity 0.85)
- Multimodal Tabular Reasoning with Privileged Structured Information (NeurIPS 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
