Evolving and Detecting Multi-Turn Deception using Geometric Signatures
Surender Suresh Kumar, Dr. M.L. Cummings · 2026
Read Evolving and Detecting Multi-Turn Deception using Geometric Signatures by Surender Suresh Kumar, Dr. M.L. Cummings (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-2605-27671.txt.
Description
Surender Suresh Kumar and Mary Cummings build a pipeline that makes multi-turn deceptive question sets and detects them with shape signals. Multi-goal genetic prompt tuning with co-evolving mutations produces real attacks, and three geometric features plus pairwise similarity stats feed a compact classifier. The detector reaches about 89 percent recall with test-time F1 from about 74 percent to about 86 percent across base, rephrased, and cut scenarios.
Catalog record
| CREATOR | Surender Suresh Kumar, Dr. M.L. Cummings |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 43008 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.27671