Whitening Reveals Cluster Commitment as the Geometric Separator of Hallucination Types
Matic Korun · 2026
Read Whitening Reveals Cluster Commitment as the Geometric Separator of Hallucination Types by Matic Korun (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-07755.txt.
Description
Korun uses PCA-whitening on GPT-2-small to separate hallucination types by clusters. A geometric taxonomy tells apart center-drift, wrong-well convergence and coverage gaps, yet prior work found the first two alike. Whitening creates a space where peak cluster match separates wrong-well outcomes from coverage gaps at Holm-corrected significance. Condition means follow the set order: wrong-well highest, center-drift middle, gaps lowest. Raising prompts per group from 15 to 30 removes a false positive in whitened entropy, and spectrum checks confirm no first-two separation.
Catalog record
| CREATOR | Matic Korun |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36354 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.07755