Articles


Papers

PaperAuthorsPublishedDOI
Semantic Tension Language (STL): A Theoretical Framework for Structured and Interpretable Knowledge RepresentationWukoNov 202510.5281/zenodo.17585432

Reports

ArticleTopicKey Finding
Context Compression for LLMsLLM MemorySTL achieves 1.76x better token efficiency than auto-compacted NL (up to 10x vs verbose prose)

Coming Soon

  • STL vs JSON for AI Applications — When structured semantics beats key-value pairs
  • Building Persistent AI Memory — Cross-session knowledge accumulation with STL
  • The Tension-Path Model — Why directional relations matter for knowledge representation