Not-Just-Scaling Laws: Towards a Better Understanding of the Downstream Impact of Language Model Design Decisions (2025.emnlp-main)
Copied to clipboard
Emmy Liu, Amanda Bertsch, Lintang Sutawika, Lindia Tjuatja, Patrick Fernandes, Lara Marinov, Michael Chen, Shreya Singhal, Carolin Lawrence, Aditi Raghunathan, Kiril Gashteovski, Graham Neubig
| Challenge: | Language model performance is largely dependent on pretraining decisions, but scaling laws based on only these two aspects do not always explain downstream task performance. |
| Approach: | They meta-analyze 92 open-source pretrained models to quantify their impact on performance. |
| Outcome: | The framework lays a foundation for more systematic investigation of how model development choices shape final capabilities. |
Similar Papers
Revisiting Scaling Laws for Language Models: The Role of Data Quality and Training Strategies (2025.acl-long)
Copied to clipboard
| Challenge: | Existing scaling laws suggest augmenting model size and training data results in enhanced performance, but recent studies reveal deviations, particularly in large language models, where performance improvements decelerate—a phenomenon known as sub-scaling. |
| Approach: | They propose a sub-optimal scaling law that better predicts performance in sub-scaling regimes by examining data quality and training strategies. |
| Outcome: | The proposed scaling law better predicts performance in sub-scaling regimes, highlighting the importance of data quality and diversity. |
Tokenizer Choice For LLM Training: Negligible or Crucial? (2024.findings-naacl)
Copied to clipboard
Mehdi Ali, Michael Fromm, Klaudia Thellmann, Richard Rutmann, Max Lübbering, Johannes Leveling, Katrin Klug, Jan Ebert, Niclas Doll, Jasper Buschhoff, Charvi Jain, Alexander Weber, Lena Jurkschat, Hammam Abdelwahab, Chelsea John, Pedro Ortiz Suarez, Malte Ostendorff, Samuel Weinbach, Rafet Sifa, Stefan Kesselheim, Nicolas Flores-Herr
| Challenge: | Recent success of large language models has been driven by curating the training dataset composition, scaling of model architectures and advancements in pretraining objectives, leaving tokenizer influence as a blind spot. |
| Approach: | They conduct a comprehensive study on the influence of tokenizer choice on LLM downstream performance by training 24 mono- and multilingual LLMs at a 2.6B parameter scale. |
| Outcome: | The proposed model can significantly impact the model's downstream performance and training costs. |
Scaling Laws for Code: Every Programming Language Matters (2026.findings-acl)
Copied to clipboard
Jian Yang, Shuyue Guo, Linzheng Chai, Wei Zhang, Aishan Liu, Chuan Hao, Zhoujun Li, Xin Zhao, Xianglong Liu, Weifeng Lv, Bryan Dai
| Challenge: | Existing studies focus on language-agnostic settings, neglecting the inherently multilingual nature of modern software development. |
| Approach: | They propose a proportion-dependent scaling law that prioritizes high-utility languages . they propose PLs to have varying effects during pre-training that affect model performance . |
| Outcome: | The proposed scaling law is based on 1000+ experiments across multiple languages and models. |
Deciphering the Impact of Pretraining Data on Large Language Models through Machine Unlearning (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing studies have suggested that the composition of the pretraining corpus exerts a significant impact upon the performance of LLMs. |
| Approach: | They analyze the impact of 48 datasets from 5 major categories of pretraining data of Large Language Models and measure their impacts on LLMs using benchmarks about nine major categories. |
| Outcome: | The proposed analysis provides insights into the organization of data to support more efficient pretraining of Large Language Models. |
A Survey of Post-Training Scaling in Large Language Models (2025.acl-long)
Copied to clipboard
Hanyu Lai, Xiao Liu, Junjie Gao, Jiale Cheng, Zehan Qi, Yifan Xu, Shuntian Yao, Dan Zhang, Jinhua Du, Zhenyu Hou, Xin Lv, Minlie Huang, Yuxiao Dong, Jie Tang
| Challenge: | Large language models (LLMs) have demonstrated proficiency in understanding and generating human natural languages. |
| Approach: | They propose a framework for scaling large language models using supervised fine-tuning, RLxF and test-time compute methodologies. |
| Outcome: | The proposed model can be used to understand and generate human natural languages. |
Data Similarity is Not Enough to Explain Language Model Performance (2023.emnlp-main)
Copied to clipboard
| Challenge: | Large language models perform well on many but not all downstream tasks. |
| Approach: | They compare large language models with downstream benchmarks to determine whether similarity measures correlate with model performance. |
| Outcome: | The results show that similarity measures are not correlated with accuracy or each other in other benchmarks. |
Scaling Parameter-Constrained Language Models with Quality Data (2024.emnlp-industry)
Copied to clipboard
Ernie Chang, Matteo Paltenghi, Yang Li, Pin-Jie Lin, Changsheng Zhao, Patrick Huber, Zechun Liu, Rastislav Rabatin, Yangyang Shi, Vikas Chandra
| Challenge: | Scaling laws in language modeling quantify training loss as a function of dataset size and model parameters, but neglect the critical role of data quality in model generalization. |
| Approach: | They propose to use effective training tokens as a combination of text diversity and syntheticity as measured by a teacher model to calculate scaling laws. |
| Outcome: | The proposed term effective training tokens is a combination of two readily-computed indicators of text diversity and syntheticity as measured by a teacher model. |
Honey, I Shrunk the Language: Language Model Behavior at Reduced Scale. (2023.findings-acl)
Copied to clipboard
| Challenge: | Recent studies have focused on high-compute settings, leaving the question of when these abilities begin to emerge largely unanswered. |
| Approach: | They investigate whether effects of pre-training can be observed when problem size is reduced, modeling a smaller, reduced-vocabulary language. |
| Outcome: | The proposed model performance is correlated with pre-training perplexity and performance. |
Training Trajectories of Language Models Across Scales (2023.acl-long)
Copied to clipboard
Mengzhou Xia, Mikel Artetxe, Chunting Zhou, Xi Victoria Lin, Ramakanth Pasunuru, Danqi Chen, Luke Zettlemoyer, Veselin Stoyanov
| Challenge: | Scaling up language models has led to unprecedented performance gains, but little is understood about how the training dynamics change as models get larger. |
| Approach: | They analyze the training checkpoints of different-sized OPT models on next-token prediction, sequence-level generation and downstream tasks. |
| Outcome: | The results show that language models of different sizes learn more during training . small models halt at hallucinations, larger ones learn to assign lower probabilities . |
Scaling Laws for Code: A More Data-Hungry Regime (2026.acl-long)
Copied to clipboard
Xianzhen Luo, Wenzhen Zheng, Qingfu Zhu, Rongyi Zhang, Houyi Li, Siming Huang, YuanTao Fan, Wanxiang Che
| Challenge: | Code Large Language Models (LLMs) are revolutionizing software engineering, but scaling laws are primarily analyzed on Natural Language (NL). |
| Approach: | They fit Chinchilla law and Farsser law to test scaling laws for code . they find code is more data-hungry and requires higher data-to-parameter ratio . |
| Outcome: | The proposed scaling laws show that the more expressive Farsser law offers greater accuracy and scales with model size. |