{"id":96,"identifier":"LOR-00096","url":"https://lifeofresearch.com/articles/96","title":"Dictionaries Without Ontologies: What a Sparse Autoencoder Latent Is Evidence For, Why Decodability, Use and Reproducibility Do Not Add Up to Canonicity, and the Identifiability Premise No Real-Model Result Has Checked","abstract":"(c) 2026 Pranay Mahendrakar. Licensed under CC BY 4.0.\nSparse autoencoders are now the default instrument for decomposing language model activations, and their latents are routinely reported as the model's features: units the network itself uses, which a wide enough dictionary has pulled out of superposition. This paper asks what a latent is evidence for. It separates four claims that the word \"feature\" runs together - that a property is linearly readable at an activation site, that the model's downstream computation depends on a direction, that the same unit recurs across training choices, and that the recurring unit is canonical, meaning unique, complete and atomic - and reads the published record against each. The record supports the first broadly, though the natural-language label attached to a latent is a weaker link than the latent itself; it supports the second unevenly, with causal roles that differ between SAE families trained on the same model; it supports the third at the grain of subspaces more readily than at the grain of individual latents; and it supports the fourth on no real language model. The reason is structural. Classical results make sparse dictionaries unique only when the data are generated by a sparse linear code meeting stated conditions; that premise has been tested only piecemeal, concept by concept, and the comparisons with a ground truth for the full dictionary that could reach canonicity have been run only on toy, synthetic, formal-language and board-game models; natural-language models have only task-specific, approximate ground truths. A short rotation argument shows how a multi-dimensional feature would produce the pattern of seed-dependent latents inside reproducible subspaces that has been reported. The paper ends with a use-indexed licensing procedure and five studies that would settle the open part. Throughout, it works from published measurements, each credited to the source that reported it.","authors":[{"name":"Pranay Mahendrakar","affiliation":"SONYTECH","orcid":"0009-0003-7224-029X"}],"type":"article","license":"CC-BY-4.0","keywords":["sparse autoencoders","mechanistic interpretability","dictionary learning","identifiability","superposition","feature universality","causal abstraction","evaluation validity"],"publication_date":"2026-10-09","version":1,"views":7,"downloads":4,"created_at":"2026-10-10 09:59:02","doi":"10.5281/zenodo.23250481","source":"Zenodo","source_url":"https://zenodo.org/records/23250481","files":[{"id":96,"name":"dictionaries-without-ontologies.pdf","size":415053,"mime":"application/pdf","downloads":4,"download_url":"https://lifeofresearch.com/articles/96/files/96/dictionaries-without-ontologies.pdf"}]}