Data Mixing Can Induce Phase Transitions in Knowledge Acquisition

📰 ArXiv cs.AI

arXiv:2505.18091v3 Announce Type: replace-cross Abstract: Large Language Models (LLMs) are typically trained on data mixtures: most data come from web scrapes, while a small portion is curated from high-quality sources with dense domain-specific knowledge. In this paper, we show that when training LLMs on such data mixtures, knowledge acquisition from knowledge-dense datasets, unlike training exclusively on knowledge-dense data (arXiv:2404.05405), does not always follow a smooth scaling law but

Published 12 May 2026

Full Article

Title: Data Mixing Can Induce Phase Transitions in Knowledge Acquisition

Abstract:
arXiv:2505.18091v3 Announce Type: replace-cross Abstract: Large Language Models (LLMs) are typically trained on data mixtures: most data come from web scrapes, while a small portion is curated from high-quality sources with dense domain-specific knowledge. In this paper, we show that when training LLMs on such data mixtures, knowledge acquisition from knowledge-dense datasets, unlike training exclusively on knowledge-dense data (arXiv:2404.05405), does not always follow a smooth scaling law but
Read full paper → ← Back to Reads