Brazilian Portuguese Agentic Text-to-SQL Distilled Trajectories

Brazilian Portuguese message-only distilled trajectories for training tool-using Text-to-SQL agents. The dataset was created by Boakpe and contains trajectories selected from LLM-judged correct conversations, preserving the agent protocol from the released code. It was last updated on June 15, 2026.

Use Cases

• Fine-tune language models for Text-to-SQL conversion based on Brazilian Portuguese natural language queries.
• Train agentic AI systems on the specific interaction protocol mentioned in the description.
• Benchmark the performance of SQL-generating agents using distilled conversation trajectories.
• Study the distillation process for tool-using agents in a non-English language context.

Strengths

• Trajectories were selected from LLM-judged correct conversations, suggesting a quality filter.
• The agent protocol used in the released code is preserved, providing consistency.
• The dataset is specifically for Brazilian Portuguese, addressing a less common language in this domain.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

huggingface, author Boakpe

Collection Method

Distilled from LLM-judged correct conversations.

Freshness

Last updated 2026-06-15 06:08:02; freshness should be verified.

Geography

Brazilian Portuguese language context.