@misc{indiciae13cf3a524193,
  title = {Depth Laws for the Precision Floor of Trained Neural Networks: Amplification, Residual Scaling, and a Quantization-Aware Training Paradox},
  author = {Ahmad S. Tarawneh},
  year = {2026},
  url = {https://arxiv.org/abs/2609.32060},
  note = {Source identifier: 2609.32060}
}