--- title: "Measuring short-form factuality in large language — Wei 2024" description: "Large language models often state false facts with confidence, and factuality is hard to measure when answers are long and full of claims." authors: ["Jason Wei", "Nguyen Karina", "Hyung Won Chung", "Yunxin Joy Jiao", "Spencer Papay", "Amelia Glaese", "John Schulman", "William Fedus"] year: 2024 venue: "arXiv:2411.04368" source: "https://arxiv.org/abs/2411.04368" doi: "10.48550/arxiv.2411.04368" arxiv: "2411.04368" licence: "arXiv — non-exclusive license to distribute" licence_url: "http://arxiv.org/licenses/nonexclusive-distrib/1.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/wei-2024-measuring-short-form-factuality-large/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/wei-2024-measuring-short-form-factuality-large/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2024](https://md.research.elamur.ai/free/en/papers/2024/page/4/) / Measuring short-form factuality in large language models # Measuring short-form factuality in large language models **Authors:** [Jason Wei](https://md.research.elamur.ai/free/en/authors/jason-wei/), Nguyen Karina, [Hyung Won Chung](https://md.research.elamur.ai/free/en/authors/hyung-won-chung/), Yunxin Joy Jiao, Spencer Papay, [Amelia Glaese](https://md.research.elamur.ai/free/en/authors/amelia-glaese/), [John Schulman](https://md.research.elamur.ai/free/en/authors/john-schulman/), William Fedus · **Year:** 2024 · **Venue:** arXiv:2411.04368 · **Licence:** [arXiv — non-exclusive license to distribute](http://arxiv.org/licenses/nonexclusive-distrib/1.0/) **Links:** [Publisher's page](https://arxiv.org/abs/2411.04368) · [PDF](https://arxiv.org/pdf/2411.04368) · [DOI 10.48550/arxiv.2411.04368](https://doi.org/10.48550/arxiv.2411.04368) · [arXiv 2411.04368](https://arxiv.org/abs/2411.04368) · [OpenAlex](https://openalex.org/W4404396608) **Citations:** 13 citations (openalex-snapshot · 2026-10-14) · [134](#cited-by) in our library, of the 140673 papers whose bibliographies we read ## Summary Large language models often state false facts with confidence, and factuality is hard to measure when answers are long and full of claims. The OpenAI authors narrow the problem to short questions that each have exactly one verifiable answer, and they release SimpleQA, a benchmark of 4,326 such questions. Human AI trainers wrote the questions adversarially: a question was kept only if at least one of four GPT-4-era model answers got it wrong. A second trainer then had to reach the same answer on their own, and the answer had to be supported by sources from two different web domains. A prompted ChatGPT grader labels each model response as correct, incorrect or not attempted. Frontier models score below 50%; the best result in the paper is 42.7% for o1-preview. The authors also use SimpleQA to test calibration. Models are more accurate when they report higher confidence or give the same answer more often, but they consistently claim more confidence than their accuracy supports. The benchmark gives an inexpensive, low-noise way to track hallucinations and whether models know the limits of their own knowledge. ## Key points - **Dataset:** 4,326 questions across many topics. The largest groups are Science & Technology (858), Politics (709) and Art (550). About a third of the answers are dates and about a quarter are people. Wikipedia is a source for roughly 3.5k of the questions. The data is released at github.com/openai/simple-evals. - **Quality check:** a third trainer re-answered 1,000 sampled questions. From the disagreements, the authors estimate that about 3% of the benchmark's reference answers are wrong. The main causes are ambiguous questions and sources that contradict each other. - **Metrics:** the paper reports overall correct, correct given attempted, and an F-score (the harmonic mean of the two). The authors note that the F-score can still reward guessing. - **Results (correct %):** o1-preview 42.7, GPT-4o 38.2, Claude-3.5-sonnet 28.9, GPT-4o-mini 8.6, o1-mini 8.1. Claude models decline to answer much more often than the others (Claude-3.5-sonnet does not attempt 35% of the questions). - **Calibration:** the larger models (o1-preview, GPT-4o) are better calibrated than their mini versions. This holds both for confidence the model states and for how often it repeats an answer across 100 samples. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* Fields: [Other](https://md.research.elamur.ai/free/en/shelves/other/) ## Cited by Cited by 134 papers of our library cite it (of the 140673 whose bibliographies we read); 10 of them have a page here, 124 not yet. - [Phi-4 Technical Report](https://md.research.elamur.ai/free/en/papers/abdin-2024-phi-4-technical-report/) - [Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities](https://md.research.elamur.ai/free/en/papers/comanici-2025-gemini-2-5-pushing-frontier/) - [Olmo 3](https://md.research.elamur.ai/free/en/papers/ettinger-2025-olmo-3/) - [GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models](https://md.research.elamur.ai/free/en/papers/glm-2025-4-glm-4-5-agentic/) - [GLM-5: from Vibe Coding to Agentic Engineering](https://md.research.elamur.ai/free/en/papers/glm-2026-5-glm-5-vibe-coding/) - [AraDynFact: Dynamic Evaluation of Factual Knowledge in Arabic](https://md.research.elamur.ai/free/en/papers/iacobacci-2026-aradynfact-dynamic-evaluation-factual-knowledge/) - [Kimi K2: Open Agentic Intelligence](https://md.research.elamur.ai/free/en/papers/kimi-2025-team-kimi-k2-open-agentic/) - [Humanity's Last Exam](https://md.research.elamur.ai/free/en/papers/phan-2025-humanity-s-last-exam/) - [Siren's Song in the AI Ocean: A Survey on Hallucination in Large Language Models](https://md.research.elamur.ai/free/en/papers/zhang-2023-siren-s-song-ai-ocean/) - [Large Language Models for Information Retrieval: A Survey](https://md.research.elamur.ai/free/en/papers/zhu-2023-large-language-models-information-retrieval/) ## Cite The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @misc{wei2024measuring, title = {{Measuring short-form factuality in large language models}}, author = {Wei, Jason and Karina, Nguyen and Chung, Hyung Won and Jiao, Yunxin Joy and Papay, Spencer and Glaese, Amelia and Schulman, John and Fedus, William}, year = {2024}, doi = {10.48550/arxiv.2411.04368}, eprint = {2411.04368}, archivePrefix = {arXiv}, url = {https://doi.org/10.48550/arxiv.2411.04368} } ``` - [← previous Measuring Multimodal Mathematical Reasoning with MATH-Vision Dataset](https://md.research.elamur.ai/free/en/papers/wang-2024-measuring-multimodal-mathematical-reasoning-math/) - [next → MedAgents: Large Language Models as Collaborators for Zero-shot Medical Reasoning](https://md.research.elamur.ai/free/en/papers/tang-2024-medagents-large-language-models-as/) - [↑ back to papers of 2024](https://md.research.elamur.ai/free/en/papers/2024/page/4/) ## References 20 references, 8 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. A. Agrawal, M. Suzgun, L. Mackey, and A. T. Kalai. Do language models know when they’re hallucinating references? Findings of EACL, 2024. URL https://arxiv.org/abs/2305.18248. [↗](https://arxiv.org/abs/2305.18248) 2. P. Anthropic. Claude 3 model card, 2024. URL [https://www-cdn.anthropic.com/de8ba9b01c9ab7cbabf5c33b80b7bbc618857627/Model_Card_Claude_3.pdf](https://www-cdn.anthropic.com/de8ba9b01c9ab7cbabf5c33b80b7bbc618857627/Model_Card_Claude_3.pdf). [↗](https://www-cdn.anthropic.com/de8ba9b01c9ab7cbabf5c33b80b7bbc618857627/Model_Card_Claude_3.pdf) 3. Q. Cheng, T. Sun, W. Zhang, S. Wang, X. Liu, M. Zhang, J. He, M. Huang, Z. Yin, K. Chen, et al. Evaluating hallucinations in chinese large language models. arXiv preprint arXiv:2310.03368, 2023. URL [https://openreview.net/forum?id=1AXvGjfF0V](https://openreview.net/forum?id=1AXvGjfF0V). — [arXiv:2310.03368](https://arxiv.org/abs/2310.03368) [↗](https://openreview.net/forum?id=1AXvGjfF0V) 4. M. Joshi, E. Choi, D. Weld, and L. Zettlemoyer. TriviaQA: A large scale distantly supervised challenge dataset for reading comprehension. In Proc of ACL, 2017. URL https://arxiv.org/abs/1705.03551. [↗](https://arxiv.org/abs/1705.03551) [our card](https://md.research.elamur.ai/free/en/papers/joshi-2017-triviaqa-large-scale-distantly-supervised/) 5. S. Kadavath, T. Conerly, A. Askell, T. Henighan, D. Drain, E. Perez, N. Schiefer, Z. Hatfield-Dodds, N. DasSarma, E. Tran-Johnson, S. Johnston, S. El-Showk, A. Jones, N. Elhage, T. Hume, A. Chen, Y. Bai, S. Bowman, S. Fort, D. Ganguli, D. Hernandez, J. Jacobson, J. Kernion, S. Kravec, L. Lovitt, K.… [find](https://scholar.google.com/scholar?q=S%2E+Kadavath%2C+T%2E+Conerly%2C+A%2E+Askell%2C+T%2E+Henighan%2C+D%2E+Drain%2C+E%2E+Perez%2C+N%2E+Schiefer%2C+Z%2E+Hatfield%2DDodds%2C+N%2E+DasSarma%2C+E%2E+Tran%2DJohnson%2C+S%2E+Johnston%2C+S%2E+El%2DShowk%2C+A%2E+Jones%2C+N%2E+Elhage%2C+T%2E+Hume%2C+A%2E+Chen%2C+Y%2E+B) 6. A. T. Kalai. Personal communication, July 2024. Communication on July 9, 2024. [find](https://scholar.google.com/scholar?q=A%2E+T%2E+Kalai%2E+Personal+communication%2C+July+2024%2E+Communication+on+July+9%2C+2024%2E) 7. S. Krishna, K. Krishna, A. Mohananey, S. Schwarcz, A. Stambler, S. Upadhyay, and M. Faruqui. Fact, fetch, and reason: A unified evaluation of retrieval-augmented generation. arXiv preprint, 2024. URL https://arxiv.org/abs/2409.12941. [↗](https://arxiv.org/abs/2409.12941) 8. T. Kwiatkowski, J. Palomaki, O. Redfield, M. Collins, A. Parikh, C. Alberti, D. Epstein, I. Polosukhin, J. Devlin, K. Lee, K. Toutanova, L. Jones, M. Kelcey, M.-W. Chang, A. M. Dai, J. Uszkoreit, Q. Le, and S. Petrov. Natural questions: A benchmark for question answering research. TACL, 2019. URL ht… [find](https://scholar.google.com/scholar?q=T%2E+Kwiatkowski%2C+J%2E+Palomaki%2C+O%2E+Redfield%2C+M%2E+Collins%2C+A%2E+Parikh%2C+C%2E+Alberti%2C+D%2E+Epstein%2C+I%2E+Polosukhin%2C+J%2E+Devlin%2C+K%2E+Lee%2C+K%2E+Toutanova%2C+L%2E+Jones%2C+M%2E+Kelcey%2C+M%2E%2DW%2E+Chang%2C+A%2E+M%2E+Dai%2C+J%2E+Uszkoreit%2C+Q%2E+L) [our card](https://md.research.elamur.ai/free/en/papers/kwiatkowski-2019-natural-questions-benchmark-question-answering/) 9. J. Li, X. Cheng, X. Zhao, J.-Y. Nie, and J.-R. Wen. HaluEval: A large-scale hallucination evaluation benchmark for large language models. In EMNLP, 2023. URL https://arxiv.org/abs/2305.11747. [↗](https://arxiv.org/abs/2305.11747) [our card](https://md.research.elamur.ai/free/en/papers/li-2023-halueval-large-scale-hallucination-evaluation/) 10. S. Lin, J. Hilton, and O. Evans. TruthfulQA: Measuring how models mimic human falsehoods. In ACL, 2022a. URL https://aclanthology.org/2022.acl-long.229. [↗](https://aclanthology.org/2022.acl-long.229) [our card](https://md.research.elamur.ai/free/en/papers/lin-2022-truthfulqa-measuring-how-models-mimic/) 11. S. Lin, J. Hilton, and O. Evans. Teaching models to express their uncertainty in words. In TMLR, 2022b. URL https://arxiv.org/abs/2205.14334. [↗](https://arxiv.org/abs/2205.14334) 12. S. Min, K. Krishna, X. Lyu, M. Lewis, W.-t. Yih, P. Koh, M. Iyyer, L. Zettlemoyer, and H. Hajishirzi. FActScore: Fine-grained atomic evaluation of factual precision in long form text generation. In EMNLP, 2023. URL https://aclanthology.org/2023. emnlp-main.741. [↗](https://aclanthology.org/2023) [our card](https://md.research.elamur.ai/free/en/papers/min-2023-factscore-fine-grained-atomic-evaluation/) 13. OpenAI. Hello gpt-4o, 2024a. URL [https://openai.com/index/hello-gpt-4o/](https://openai.com/index/hello-gpt-4o/). [↗](https://openai.com/index/hello-gpt-4o/) 14. OpenAI. Openai o1-mini, 2024b. URL [https://openai.com/index/openai-o1-mini-advancing-cost-efficient-reasoning/](https://openai.com/index/openai-o1-mini-advancing-cost-efficient-reasoning/). [↗](https://openai.com/index/openai-o1-mini-advancing-cost-efficient-reasoning/) 15. OpenAI. Learning to reason with llms, 2024c. learning-to-reason-with-llms/. [find](https://scholar.google.com/scholar?q=OpenAI%2E+Learning+to+reason+with+llms%2C+2024c%2E+learning%2Dto%2Dreason%2Dwith%2Dllms%2F%2E) 16. T. Vu, M. Iyyer, X. Wang, N. Constant, J. Wei, J. Wei, C. Tar, Y.-H. Sung, D. Zhou, Q. Le, and T. Luong. Freshllms: Refreshing large language models with search engine augmentation, 2023. URL https://arxiv.org/abs/2310.03214. [↗](https://arxiv.org/abs/2310.03214) [our card](https://md.research.elamur.ai/free/en/papers/vu-2023-freshllms-refreshing-large-language-models/) 17. X. Wang, J. Wei, D. Schuurmans, Q. Le, E. Chi, S. Narang, A. Chowdhery, and D. Zhou. Self-consistency improves chain of thought reasoning in language models. In ICML, 2023. URL https://arxiv.org/abs/2203.11171. [↗](https://arxiv.org/abs/2203.11171) [our card](https://md.research.elamur.ai/free/en/papers/wang-2022-self-consistency-improves-chain-thought/) 18. J. Wei, C. Yang, X. Song, Y. Lu, N. Hu, J. Huang, D. Tran, D. Peng, R. Liu, D. Huang, C. Du, and Q. V. Le. Long-form factuality in large language models. In NeurIPS, 2024. URL https://arxiv.org/abs/2403.18802. [↗](https://arxiv.org/abs/2403.18802) [our card](https://md.research.elamur.ai/free/en/papers/wei-2024-long-form-factuality-large-language/) 19. Y. Zhao, J. Zhang, I. Chern, S. Gao, P. Liu, J. He, et al. FELM: Benchmarking factuality evaluation of large language models. NeurIPS, 2024. URL https://arxiv.org/abs/2310.00741. [↗](https://arxiv.org/abs/2310.00741) 20. URL [https://openai.com/index/](https://openai.com/index/) [↗](https://openai.com/index/)