A Best Match Is Not a Twin: What Feature Similarity Between State Space Models and Transformers Certifies, Why the Null and the Shared Inputs Decide the Verdict, and the Controls That Would Show Two Architectures Learned the Same Thing

APA

Mahendrakar, P. (2026). A Best Match Is Not a Twin: What Feature Similarity Between State Space Models and Transformers Certifies, Why the Null and the Shared Inputs Decide the Verdict, and the Controls That Would Show Two Architectures Learned the Same Thing [Preprint]. Life of Research. https://doi.org/10.5281/zenodo.23652519

Copy APA

BibTeX

@misc{lor1422026,
  title        = {A Best Match Is Not a Twin: What Feature Similarity Between State Space Models and Transformers Certifies, Why the Null and the Shared Inputs Decide the Verdict, and the Controls That Would Show Two Architectures Learned the Same Thing},
  author       = {Pranay Mahendrakar},
  year         = {2026},
  publisher    = {Life of Research},
  doi          = {10.5281/zenodo.23652519},
  url          = {https://lifeofresearch.com/preprints/142},
  keywords     = {universality hypothesis, representational similarity, sparse autoencoders, state space models, Mamba, mechanistic interpretability, null calibration, Platonic representation hypothesis},
  note         = {LOR-00142}
}

Copy BibTeX More on the record page