Liu et al. (2025h) introduce the concept of superposition, where models represent more features than the dimensions they have, and propose a power-law loss with model size without assuming power laws in data.

Authors

• Person: Not available Organization: arXiv
  A Survey on the Theory and Mechanism of Large Language Models

Sources

• A Survey on the Theory and Mechanism of Large Language Models arxiv.org arXiv via serper

Referenced by nodes (1)

• Superposition concept