LLM Fingerprinting
Perplexity-Based Detection of Single-Model LLM Output
A reference model's perplexity score reveals whether a document was machine-generated.
Marcus Elesin · October 4, 2026 · 11 min read
Contributing Researcher, Benchmarks
Marcus Elesin
3 stories
LLM Fingerprinting
Statistical Fingerprinting Methods Across Major Language Models
Every language model leaves a detectable statistical signature in its token choices.
Marcus Elesin · October 3, 2026 · 11 min read
Watermark Detection & Evasion
Detection Evasion Risks When Using a Single LLM Repeatedly
Repeated use of the same LLM compounds detection risk across your entire body of work.
Marcus Elesin · September 25, 2026 · 10 min read