TRM_Genesis_RealWorld_40k is a dataset of 40,000 fact-based question-and-answer examples created by Within Us AI. Each answer includes citations (fact_sources) and a verification tag, while prompts are designed to be short and optimized for tiny contexts. The dataset was last updated on January 4, 2026.
Use Cases
- Training tiny models for factual accuracy based on the dataset's cited answers.
- Benchmarking model performance on constrained-context prompts as described.
- Evaluating citation and source verification capabilities using the fact_sources field.
- Developing knowledge verification systems based on the dataset's verification tags.
Strengths
- Contains 40,000 examples, providing a substantial corpus for training.
- Each answer includes fact_sources (citations) and a verification tag, supporting verifiable accuracy.
- Prompts are described as TRM-native, meaning they are short and optimized for tiny contexts.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is known, but other structural details like file formats and license are unknown.
- Data may reflect bias inherent to the sources used for the factual Q&A.
Provenance
- Source
- WithinUsAI
- Freshness
- Last updated 2026-01-04 00:10:29; freshness should be verified.