AI RESEARCH
Polysemanticity or Polysemy? Lexical Identity Confounds Superposition Metrics
arXiv CS.AI
•
ArXi:2604.00443v1 Announce Type: cross If the same neuron activates for both "lender" and "riverside," standard metrics attribute the overlap to superposition--the neuron must be compressing two unrelated concepts. This work explores how much of the overlap is due a lexical confound: neurons fire for a shared word form (such as "bank") rather than for two compressed concepts.