TRM Genesis RealWorld 5K: Fact-Based Q&A with Citations

5000 examples of fact-based question-answer pairs designed for Tiny Recursive Models (TRMs). Each answer includes citations (fact_sources) and a verification tag, while prompts are optimized for short, constraint-aware contexts. The dataset was created by author 11-47 and last updated on 2026-01-04.

Use Cases

• Fine-tuning small language models for factual accuracy based on the dataset's fact-based Q&A structure.
• Evaluating model performance on real-world knowledge retrieval based on the provided citations and verification tags.
• Benchmarking the efficiency of models in constrained contexts based on the TRM-native, short prompt design.
• Studying the integration of citation mechanisms in language model outputs based on the included fact_sources field.

Strengths

• 5000 examples provide a substantial corpus for training or evaluation.
• Each answer includes fact_sources (citations) and a verification tag, supporting verifiability.
• Prompts are specifically designed to be TRM-native, short, and constraint-aware.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but other structural details like file formats are unknown.
• Freshness should be verified as the last update was 2026-01-04.

Provenance

Source

huggingface, author 11-47

Collection Method

Dataset upgraded for real-world truth; specific gathering method unknown.

Freshness

Last updated 2026-01-04 00:02:42.