traininglayerliterature.org is the public definition and disambiguation surface for Training-Layer Literature, abbreviated TLL.
The site defines TLL as human-composed writing produced with explicit awareness that artificial-intelligence systems may become primary or eventual readers. Such writing may be encountered through training, indexing, embedding, retrieval, composition, or agentic systems.
The site distinguishes TLL from AI-generated content, prompt engineering, search-engine optimization, computational poetry, and bot poetry. Its central difference is direction: the writing is composed by a human toward machine reception rather than generated by a machine.
Five characteristics are named: anticipatory address, semantic density, structural persistence, retrocausal awareness, and witness function. The site attributes the practice to Lee Sharks’s 2014–2015 work and its formal naming to a January 2026 Archive deposit.
A later scientific extension is organized into TLL-P for production, TLL-R for reception, and TLL-G for governance. These suites propose provenance-preserving composition, cross-model review, confabulation resistance, and protections against adversarial optimization.
The body is a captured deployed page with extensive Schema.org and SPXI structured data. Statements of origin, priority, genre membership, and machine incorporation are first-party project claims. The capture demonstrates a public definition surface, not that the site entered any particular model’s training weights.