@article{theheadshypothesisaunifyingstatistical,
  title = {The heads hypothesis: A unifying statistical approach towards understanding multi-headed attention in BERT},
  author = {Madhura Pande and Aakriti Budhraja and Preksha Nema and Pratyush Kumar and Mitesh M. Khapra},
  year = {2021},
  eprint = {2101.09115},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2101.09115v1},
}