Performance Law of Large Language Models
This work addresses the practical need for developers to efficiently allocate computational resources and choose LLM architectures without extensive experiments, though it is incremental as it builds on existing scaling law concepts.
The authors tackled the problem of predicting real-world performance of large language models (LLMs) beyond loss estimation by developing an empirical 'Performance Law' equation that accurately predicts MMLU scores based on key hyperparameters and training data size, achieving high accuracy across diverse models.
Guided by the belief of the scaling law, large language models (LLMs) have achieved impressive performance in recent years. However, scaling law only gives a qualitative estimation of loss, which is influenced by various factors such as model architectures, data distributions, tokenizers, and computation precision. Thus, estimating the real performance of LLMs with different training settings rather than loss may be quite useful in practical development. In this article, we present an empirical equation named "Performance Law" to directly predict the MMLU score of an LLM, which is a widely used metric to indicate the general capability of LLMs in real-world conversations and applications. Based on only a few key hyperparameters of the LLM architecture and the size of training data, we obtain a quite accurate MMLU prediction of various LLMs with diverse sizes and architectures developed by different organizations in different years. Performance law can be used to guide the choice of LLM architecture and the effective allocation of computational resources without extensive experiments.