@misc{indiciae7c2d6d84d247,
  title = {NeuronMLP: Efficient LLM Inference via Singular Value Decomposition Compression and Tiling on AWS Trainium},
  author = {Dinghong Song and Jierui Xu and Weichu Yang and Pengfei Su and Dong Li},
  year = {2026},
  url = {https://arxiv.org/abs/2510.25977},
  note = {Source identifier: 2510.25977}
}