s-emanuilov/rivers-qa

Dataset for Evaluating Hallucination Mitigation in Language Models

This dataset contains structured question-answer pairs derived from U.S. river knowledge, designed to test factual grounding and epistemic discipline in LLMs. The dataset includes raw data for 9,538 river entities with 21 attributes (length, discharge, source/mouth locations, etc.) and an evaluation set of 17,726 question-answer pairs targeting factual relationships.

Citation

@article{ackermann2025stemming,
  title={Stemming Hallucination in Language Models Using a Licensing Oracle},
  author={Ackermann, Richard and Emanuilov, Simeon},
  year={2025}
}