@misc{arxiv_2609_14174v1,
  title = {Inherited Heads: Audio language models track speakers with their text backbone's attention, and an attention-mass ranking retrieves a different set},
  author = {{Bojro Das}},
  eprint = {2609.14174v1},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.14174v1}
}