Alpaca Guarani Taco: Multilingual Instruction-Response Pairs

Saillab's dataset for the TaCo paper contains instruction-response pairs formatted for instruction tuning. The data likely includes text in Guarani and English, structured with instruction, input, and output fields. It was last updated on September 20, 2024.

Use Cases

• Instruction tuning of language models based on the structured 'instruction'-'input'-'output' format
• Training or evaluating machine translation systems between Guarani and English based on the parallel response structure
• Studying cross-lingual instruction following capabilities in low-resource language contexts

Strengths

• Follows a structured format for instruction tuning as outlined in a referenced academic paper
• Provides parallel text in Guarani and English, supporting bilingual tasks
• Last updated on 2024-09-20, indicating recent maintenance

Limitations

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

Provenance

Source

saillab

Collection Method

Created for the TaCo paper; exact collection method is unspecified.

Freshness

Last updated 2024-09-20 22:08:23; freshness should be verified