Scaling Parameter-Constrained Language Models with Quality Data

Ernie Chang, Matteo Paltenghi, Yang Li, Pin-Jie Lin, Changsheng Zhao, Patrick Huber, Zechun Liu, Rastislav Rabatin, Yangyang Shi, Vikas Chandra


Abstract
Scaling laws in language modeling traditionally quantify training loss as a function of dataset size and model parameters, providing compute-optimal estimates but often neglecting the impact of data quality on model generalization.In this paper, we extend the conventional understanding of scaling law by offering a microscopic view of data quality within the original formulation – effective training tokens – which we posit to be a critical determinant of performance for parameter-constrained language models.Specifically, we formulate the proposed term of effective training tokens to be a combination of two readily-computed indicators of text:(i) text diversity and (ii) syntheticity as measured by a teacher model.We pretrained over 200 models of 25M to 1.5B parameters on a diverse set of sampled, synthetic data, and estimated the constants that relate text quality, model size, training tokens, and eight reasoning task accuracy scores.We demonstrated the estimated constants yield +0.83 Pearson correlation with true accuracies, and analyze it in scenarios involving widely-used data techniques such as data sampling and synthesis which aim to improve data quality.
Anthology ID:
2024.emnlp-industry.8
Volume:
Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing: Industry Track
Month:
November
Year:
2024
Address:
Miami, Florida, US
Editors:
Franck Dernoncourt, Daniel Preoţiuc-Pietro, Anastasia Shimorina
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
80–97
Language:
URL:
https://aclanthology.org/2024.emnlp-industry.8/
DOI:
10.18653/v1/2024.emnlp-industry.8
Bibkey:
Cite (ACL):
Ernie Chang, Matteo Paltenghi, Yang Li, Pin-Jie Lin, Changsheng Zhao, Patrick Huber, Zechun Liu, Rastislav Rabatin, Yangyang Shi, and Vikas Chandra. 2024. Scaling Parameter-Constrained Language Models with Quality Data. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing: Industry Track, pages 80–97, Miami, Florida, US. Association for Computational Linguistics.
Cite (Informal):
Scaling Parameter-Constrained Language Models with Quality Data (Chang et al., EMNLP 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.emnlp-industry.8.pdf