Nemotron RL Instruction Following Structured Outputs V2: Model Evaluation Tasks

Nemotron RL Instruction Following Structured Outputs V2 is a dataset for evaluating large language models on structured output generation. It was created by NVIDIA and last updated on June 4, 2026. The dataset includes two splits testing capabilities like freeform text generation and diversified tasks across multiple data formats.

Use Cases

• Benchmarking model performance on freeform text generation based on JSON, YAML, and XML schemas.
• Evaluating model capability for direct extraction of structured data from documents.
• Testing translation accuracy between different structured data formats like JSON and TOML.
• Assessing a model's ability for multistep translation from known data into a target format.
• Measuring performance on schema-only generation tasks without example data.

Strengths

• Includes two distinct test splits for varied evaluation scenarios.
• Covers multiple structured output formats including JSON, YAML, XML, TOML, and CSV.
• Created by NVIDIA, a leading organization in AI research and development.

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

NVIDIA

Collection Method

Likely synthetically generated for model evaluation.

Freshness

Last updated 2026-06-04 05:11:45; freshness should be verified.