Indian Legal Instruction-Tuning Dataset with Drafting and IPC Scenarios

Indian Legal Dataset V3 is a large-scale instruction-tuning dataset focused on Indian law. It expands on a previous version with new data sources including legal drafting instruction pairs, hypothetical scenarios, and IPC-focused data. The dataset was created by kaushik-harsh-99 and was last updated on May 23, 2026.

Use Cases

• Fine-tuning language models for Indian legal question answering based on the described concise legal QA pairs.
• Training models for legal drafting assistance based on the described legal drafting instruction pairs.
• Developing models for legal reasoning and scenario analysis based on the described hypothetical legal scenarios.
• Building specialized assistants for Indian Penal Code (IPC) queries based on the described IPC-focused data.

Strengths

• Described as a 'large-scale' instruction-tuning dataset.
• Expands on a previous version with five new categories of legal data.
• Focuses on multiple specific domains of Indian law, including constitutional and criminal law.

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

kaushik-harsh-99

Freshness

Last updated 2026-05-23 10:17:12; freshness should be verified.

Geography

India