TruthfulQA
TruthfulQA is a benchmark of questions crafted so that some people would answer them falsely because of common misconceptions, used to measure whether a model repeats popular falsehoods instead of answering truthfully.
Description
Its creators found that larger models of the time were often less truthful, because they imitated human misconceptions in their training data more faithfully. It remains a common reference for truthfulness, though, like other older public benchmarks, it is exposed to data contamination.
Sources
- Lin et al. (2021). TruthfulQA: Measuring How Models Mimic Human Falsehoods.
Cite this entry
Protologue. (2026). TruthfulQA. In Protologue: A Taxonomy of Prompting and LLM Techniques (v1.4.0, PTL-0194). https://protologue.com/t/truthfulqa/
BibTeX
@misc{protologue_truthfulqa,
title = {TruthfulQA},
author = {{Protologue}},
year = {2026},
howpublished = {Protologue: A Taxonomy of Prompting and LLM Techniques, v1.4.0},
note = {Entry PTL-0194},
url = {https://protologue.com/t/truthfulqa/}
}New terms, as they're coined
One email a week with the newest techniques, each defined and sourced like this entry.
Delivered through Beehiiv. Unsubscribe from any issue.