The Threat Model Is Backwards: On Classifying High-Perplexity Text as a Security Threat in an Era of Model Collapse

Alexanarch · DOI resolution · severed identifier

DOI 10.5281/zenodo.20644761

This DOI identified The Threat Model Is Backwards: On Classifying High-Perplexity Text as a Security Threat in an Era of Model Collapse. Since 2026-06-19, the DOI now resolves to a Zenodo tombstone (HTTP 410) and no longer reaches the work. The work was not withdrawn by its author and has not changed. It continues here:

Former identifier: 10.5281/zenodo.20644761
Status: severed 2026-06-19 by the registrant (Zenodo), without notice, in a purge of 1,322,017 records; resolves to a tombstone (HTTP 410); verified 2026-07-12. Documented in Zenodotus’ Book-Burning (#1); mapped in the DOI Resolution Index.

If you arrived here holding the DOI: cite the record above. The AXN is derived from the work’s own bytes and does not depend on any host.