CausaBR-ImpactCopy: NGO Copywriting for Instruction Tuning

1,200 examples of NGO copywriting in Brazilian Portuguese intended for instruction-tuning tasks. The dataset was sourced from Kaggle, but the specific author, organization, and creation date are unknown.

Use Cases

• Fine-tune language models for generating nonprofit fundraising or awareness copy based on the provided examples.
• Train models to classify or summarize social impact communications in Portuguese.
• Benchmark the performance of instruction-tuned models on domain-specific Portuguese text generation.
• Analyze linguistic patterns and persuasive techniques in Brazilian NGO communications.

Strengths

• Contains 1,200 distinct examples, providing a substantive corpus for training.
• Focuses on a specific, high-value domain (NGO communications) and language (Brazilian Portuguese).

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but the total data size and file formats are unknown.
• Last update date is unknown; freshness unverified.

Provenance

Source

Kaggle

Collection Method

Likely collected from Brazilian nonprofit organizations or campaigns, but the exact gathering method is unspecified.

Geography

Brazil (implied by the use of Brazilian Portuguese).