@misc{indiciae487028155745,
  title = {The Cost of Down-Scaling Language Models: Fact Recall Deteriorates before In-Context Learning},
  author = {Tian Jin and Nolan Clement and Xin Dong and Vaishnavh Nagarajan and Michael Carbin and Jonathan Ragan-Kelley and Gintare Karolina Dziugaite},
  year = {2023},
  url = {https://arxiv.org/abs/2310.04680},
  note = {Source identifier: 2310.04680}
}