On linear representations and pretraining data frequency in language models

{
  "doi": null,
  "year": 2025,
  "title": "On linear representations and pretraining data frequency in language models",
  "venue": null,
  "authors": [
    "Jack Merullo",
    "Noah A. Smith",
    "Sarah Wiegreffe",
    "Yanai Elazar"
  ],
  "arxiv_id": "2504.12459",
  "s2_paper_id": null,
  "ingest_status": "referenced-only"
}