Atos De Fala Ptbr: Brazilian Portuguese Speech Acts with Span-Level Annotations

An annotated dataset of speech acts for Brazilian Portuguese, where each example is a sentence with one or more marked spans. Each span receives a speech act label based on pragmatic theory, aligning with the ISO 24617-2 dialogue acts inventory. The dataset was created by lucianfialho and was last updated on 2026-06 06.

Use Cases

• Training models for speech act recognition based on span-level annotations.
• Benchmarking dialogue systems on pragmatic understanding tasks for Brazilian Portuguese.
• Analyzing linguistic patterns in Brazilian Portuguese based on the annotated speech acts.
• Developing tools for conversational AI that require intent and dialogue act classification.

Strengths

• Annotations are grounded in pragmatic theory, specifically Searle's speech acts aligned to ISO 24617-2.
• Data is annotated at the span level, providing granular labels for parts of sentences.
• The dataset focuses on Brazilian Portuguese, a less-resourced language for such tasks.

Limitations

• Row count, column definitions, and file size are unknown, limiting suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data may reflect bias inherent to its specific collection and annotation process on Hugging Face.

Provenance

Source

lucianfialho on Hugging Face.

Collection Method

Likely manually or semi-automatically annotated based on theoretical framework.

Time Range

null

Freshness

Last updated 2026-06-06 15:45:54; freshness should be verified.

Geography

Brazilian Portuguese language context.