@misc{indiciae76a1e3c14c34,
  title = {Multi-layer Cross-attention is Provably Optimal for Multi-modal In-context Learning},
  author = {Nicholas Barnfield and Subhabrata Sen and Pragya Sur},
  year = {2026},
  url = {https://arxiv.org/abs/2602.04872},
  note = {Source identifier: 2602.04872}
}