Dissecting LLM Regurgitation with the infini-gram Engine

TuhinChakr · x · 2026-08-01

Researchers from Allen AI and Stony Brook University are using the open-source infini-gram engine to dissect AI-generated prose.

The study investigates whether model-generated text is genuinely original or exact matches (regurgitation) from its training data. This approach reveals the mechanics of LLM text generation beyond what standard AI-writing detectors can capture.

Related event: Ai2's infini-gram Engine Tracks AI Writing Plagiarism(9 posts)→

Original post →

More from Research

Research channel →