Position: Language model developers should report train-test overlap

Andy K Zhang, Kevin Klyman, Yifan Mai, Yoav Levine, Yian Zhang, Rishi Bommasani, Percy Liang
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:82460-82482, 2025.

Abstract

Language models are extensively evaluated, but correctly interpreting evaluation results requires knowledge of train-test overlap, which refers to the extent to which the language model is trained on the very data it is being tested on. The public currently lacks adequate information about train-test overlap: most models have no public train-test overlap statistics, and third parties cannot directly measure train-test overlap since they do not have access to the training data. To make this clear, we document the practices of 30 models, finding that just 9 models report train-test overlap: 4 models release training data under open-source licenses, enabling the community to directly measure train-test overlap, and 5 models publish their train-test overlap methodology and statistics. By engaging with language model developers, we provide novel information about train-test overlap for three additional models. Overall, this position paper argues that language model developers should publish train-test overlap statistics and/or training data whenever they report evaluation results on public test sets. We hope our work increases transparency into train-test overlap to increase the community-wide trust in model evaluations.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-zhang25dt, title = {Position: Language model developers should report train-test overlap}, author = {Zhang, Andy K and Klyman, Kevin and Mai, Yifan and Levine, Yoav and Zhang, Yian and Bommasani, Rishi and Liang, Percy}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {82460--82482}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/zhang25dt/zhang25dt.pdf}, url = {https://proceedings.mlr.press/v267/zhang25dt.html}, abstract = {Language models are extensively evaluated, but correctly interpreting evaluation results requires knowledge of train-test overlap, which refers to the extent to which the language model is trained on the very data it is being tested on. The public currently lacks adequate information about train-test overlap: most models have no public train-test overlap statistics, and third parties cannot directly measure train-test overlap since they do not have access to the training data. To make this clear, we document the practices of 30 models, finding that just 9 models report train-test overlap: 4 models release training data under open-source licenses, enabling the community to directly measure train-test overlap, and 5 models publish their train-test overlap methodology and statistics. By engaging with language model developers, we provide novel information about train-test overlap for three additional models. Overall, this position paper argues that language model developers should publish train-test overlap statistics and/or training data whenever they report evaluation results on public test sets. We hope our work increases transparency into train-test overlap to increase the community-wide trust in model evaluations.} }
Endnote
%0 Conference Paper %T Position: Language model developers should report train-test overlap %A Andy K Zhang %A Kevin Klyman %A Yifan Mai %A Yoav Levine %A Yian Zhang %A Rishi Bommasani %A Percy Liang %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-zhang25dt %I PMLR %P 82460--82482 %U https://proceedings.mlr.press/v267/zhang25dt.html %V 267 %X Language models are extensively evaluated, but correctly interpreting evaluation results requires knowledge of train-test overlap, which refers to the extent to which the language model is trained on the very data it is being tested on. The public currently lacks adequate information about train-test overlap: most models have no public train-test overlap statistics, and third parties cannot directly measure train-test overlap since they do not have access to the training data. To make this clear, we document the practices of 30 models, finding that just 9 models report train-test overlap: 4 models release training data under open-source licenses, enabling the community to directly measure train-test overlap, and 5 models publish their train-test overlap methodology and statistics. By engaging with language model developers, we provide novel information about train-test overlap for three additional models. Overall, this position paper argues that language model developers should publish train-test overlap statistics and/or training data whenever they report evaluation results on public test sets. We hope our work increases transparency into train-test overlap to increase the community-wide trust in model evaluations.
APA
Zhang, A.K., Klyman, K., Mai, Y., Levine, Y., Zhang, Y., Bommasani, R. & Liang, P.. (2025). Position: Language model developers should report train-test overlap. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:82460-82482 Available from https://proceedings.mlr.press/v267/zhang25dt.html.

Related Material