Research map / Large language models / Agents, code and math
SheetCopilot: Bringing Software Productivity to the Next Level through Large Language Models
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use SheetCopilot instead of generating and running VBA code | Spreadsheet task control · SheetCopilot benchmark, 100% | Pass@1: 44.3% vs. 16.3% | Helped |
| Add sheet-state feedback to the planner (ablation) | SheetCopilot · SheetCopilot benchmark, full dataset | Pass@1: 18.6% -> 24.4% | Helped |
| Add error feedback to the planner (ablation) | SheetCopilot · SheetCopilot benchmark, full dataset | Exec@1: 67.9% -> 92.8% | Helped |
| Add both sheet-state and error feedback (ablation) | SheetCopilot · SheetCopilot benchmark, full dataset | Pass@1: 18.6% -> 44.3% | Helped |
| Add both feedback types (ablation) | SheetCopilot · SheetCopilot benchmark, full dataset | A50: 1.00 -> 1.50 | Mixed |
| Add external atomic-action documentation and usage examples (ablation) | SheetCopilot · SheetCopilot benchmark, full dataset | Exec@1: 56.6% -> 67.9% | Helped |
| Add external documentation, without usage examples (ablation) | SheetCopilot with state and error feedback · SheetCopilot benchmark, full dataset | Pass@1: 18.1% -> 28.1% | Helped |
| Add usage examples to the external documentation (ablation) | SheetCopilot with state and error feedback · SheetCopilot benchmark, full dataset | Pass@1: 28.1% -> 44.3% | Helped |
| Replace official atomic-action names with distant synonyms (ablation) | SheetCopilot · SheetCopilot benchmark, full dataset | Pass@1: 44.3% -> 45.2% | Helped |
| Replace official atomic-action names with distant synonyms (ablation) | SheetCopilot · SheetCopilot benchmark, full dataset | Exec@1: 87.3% -> 86.9% | Hurt |
| Integrate chart-property setting into CreateChart (ablation) | SheetCopilot · Chart-related tasks | Pass@1: 43.3% -> 37.2% | Hurt |
| Integrate chart-property setting into CreateChart (ablation) | SheetCopilot · Chart-related tasks | A50: 1.25 -> 1.00 | Helped |
Closest papers by meaning
- BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructions (ICLR 2025 · similarity 0.87)
- SpreadsheetBench: Towards Challenging Real World Spreadsheet Manipulation (NeurIPS 2024 · similarity 0.87)
- ProgramBench: Can Language Models Rebuild Programs From Scratch? (NeurIPS 2026 · similarity 0.85)
- CRAFT: Customizing LLMs by Creating and Retrieving from Specialized Toolsets (ICLR 2024 · similarity 0.85)
- Large Language Models as Tool Makers (ICLR 2024 · similarity 0.85)
- FLAME: A Small Language Model for Spreadsheet Formulas (AAAI 2024 · 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.