A Design Space for the Critical Validation of LLM-Generated Tabular Data
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866916724919500800 |
|---|---|
| author | Sachdeva, Madhav Narayanan, Christopher Wiedenkeller, Marvin Sedlakova, Jana Bernard, Jürgen |
| author_facet | Sachdeva, Madhav Narayanan, Christopher Wiedenkeller, Marvin Sedlakova, Jana Bernard, Jürgen |
| contents | LLM-generated tabular data is creating new opportunities for data-driven applications in academia, business, and society. To leverage benefits like missing value imputation, labeling, and enrichment with context-aware attributes, LLM-generated data needs a critical validation process. The number of pioneering approaches is increasing fast, opening a promising validation space that, so far, remains unstructured. We present a design space for the critical validation of LLM-generated tabular data with two dimensions: First, the Analysis Granularity dimension: from within-attribute (single-item and multi-item) to across-attribute perspectives (1 x 1, 1 x m, and n x n). Second, the Data Source dimension: differentiating between LLM-generated values, ground truth values, explanations, and their combinations. We discuss analysis tasks for each dimension cross-cut, map 19 existing validation approaches, and discuss the characteristics of two approaches in detail, demonstrating descriptive power. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_04487 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A Design Space for the Critical Validation of LLM-Generated Tabular Data Sachdeva, Madhav Narayanan, Christopher Wiedenkeller, Marvin Sedlakova, Jana Bernard, Jürgen Human-Computer Interaction LLM-generated tabular data is creating new opportunities for data-driven applications in academia, business, and society. To leverage benefits like missing value imputation, labeling, and enrichment with context-aware attributes, LLM-generated data needs a critical validation process. The number of pioneering approaches is increasing fast, opening a promising validation space that, so far, remains unstructured. We present a design space for the critical validation of LLM-generated tabular data with two dimensions: First, the Analysis Granularity dimension: from within-attribute (single-item and multi-item) to across-attribute perspectives (1 x 1, 1 x m, and n x n). Second, the Data Source dimension: differentiating between LLM-generated values, ground truth values, explanations, and their combinations. We discuss analysis tasks for each dimension cross-cut, map 19 existing validation approaches, and discuss the characteristics of two approaches in detail, demonstrating descriptive power. |
| title | A Design Space for the Critical Validation of LLM-Generated Tabular Data |
| topic | Human-Computer Interaction |
| url | https://arxiv.org/abs/2505.04487 |