Findings of the BabyLM Challenge: Sample-Efficient Pretraining on Developmentally Plausible Corpora
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916685460537344 |
|---|---|
| author | Warstadt, Alex Mueller, Aaron Choshen, Leshem Wilcox, Ethan Zhuang, Chengxu Ciro, Juan Mosquera, Rafael Paranjape, Bhargavi Williams, Adina Linzen, Tal Cotterell, Ryan |
| author_facet | Warstadt, Alex Mueller, Aaron Choshen, Leshem Wilcox, Ethan Zhuang, Chengxu Ciro, Juan Mosquera, Rafael Paranjape, Bhargavi Williams, Adina Linzen, Tal Cotterell, Ryan |
| contents | Children can acquire language from less than 100 million words of input. Large language models are far less data-efficient: they typically require 3 or 4 orders of magnitude more data and still do not perform as well as humans on many evaluations. These intensive resource demands limit the ability of researchers to train new models and use existing models as developmentally plausible cognitive models. The BabyLM Challenge is a communal effort in which participants compete to optimize language model training on a fixed data budget. Submissions are compared on various evaluation tasks targeting grammatical ability, downstream task performance, and generalization. Participants can submit to up to three tracks with progressively looser data restrictions. From over 30 submissions, we extract concrete recommendations on how best to train data-efficient language models, and on where future efforts should (and perhaps should not) focus. The winning submissions using the LTG-BERT architecture (Samuel et al., 2023) outperformed models trained on trillions of words. Other submissions achieved strong results through training on shorter input sequences or training a student model on a pretrained teacher. Curriculum learning attempts, which accounted for a large number of submissions, were largely unsuccessful, though some showed modest improvements. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_08165 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Findings of the BabyLM Challenge: Sample-Efficient Pretraining on Developmentally Plausible Corpora Warstadt, Alex Mueller, Aaron Choshen, Leshem Wilcox, Ethan Zhuang, Chengxu Ciro, Juan Mosquera, Rafael Paranjape, Bhargavi Williams, Adina Linzen, Tal Cotterell, Ryan Computation and Language Children can acquire language from less than 100 million words of input. Large language models are far less data-efficient: they typically require 3 or 4 orders of magnitude more data and still do not perform as well as humans on many evaluations. These intensive resource demands limit the ability of researchers to train new models and use existing models as developmentally plausible cognitive models. The BabyLM Challenge is a communal effort in which participants compete to optimize language model training on a fixed data budget. Submissions are compared on various evaluation tasks targeting grammatical ability, downstream task performance, and generalization. Participants can submit to up to three tracks with progressively looser data restrictions. From over 30 submissions, we extract concrete recommendations on how best to train data-efficient language models, and on where future efforts should (and perhaps should not) focus. The winning submissions using the LTG-BERT architecture (Samuel et al., 2023) outperformed models trained on trillions of words. Other submissions achieved strong results through training on shorter input sequences or training a student model on a pretrained teacher. Curriculum learning attempts, which accounted for a large number of submissions, were largely unsuccessful, though some showed modest improvements. |
| title | Findings of the BabyLM Challenge: Sample-Efficient Pretraining on Developmentally Plausible Corpora |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2504.08165 |