SAB-Bench: Social Attribution Benchmark with Vignettes

SAB-Bench is a vignette-based benchmark dataset for studying social attribution, the reasoning process of attributing causes and assigning responsibility in social interactions. The dataset was created by author yzx010 and was last updated on Hugging Face on April 21, 2026. The specific scale, row count, and column details are not provided in the available metadata.

Use Cases

• Benchmarking AI models on social attribution tasks based on vignette descriptions.
• Studying causal reasoning in social interactions based on the described vignette structure.
• Evaluating model performance on assigning responsibility or blame in described social scenarios.
• Training models to infer agents' intentions and causes from described social events.

Strengths

• Dataset is explicitly designed as a benchmark for the specific research area of social attribution.
• The description provides a clear, formal definition of the core task (attributing causes and assigning responsibility in social interactions).
• The dataset was last updated on April 21, 2026, indicating recent maintenance.

Limitations

• Row count, column definitions, and file formats are unknown, which limits suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The license is unspecified, which may restrict permissible use cases.

Provenance

Source

yzx010 on Hugging Face

Freshness

Last updated 2026-04-21 10:34:39.