The Structure and Mathematical Model of Token Cosmos
February 18, 2026 • Jerry Zhang
The Structure and Mathematical Model of Token Cosmos
Read MoreResearch Blog
Latest research updates, technical insights, and explorations into AGI emergence through semantic traversals.
Differential Topological Properties and Discretization Limits of High-Dimensional Semantic Manifolds
February 04, 2026 • Jerry Zhang
Based on experimental data from OT-SGN v45.1, this paper explores the topological evolution of semantic manifolds, defining the 'Semantic Planck Length' and identifying the boundaries of LLM interp...
Read MoreOT-SGN: Geometric Navigation and Control of Large Language Model Hallucinations via Optimal Transport and Manifold Tunneling
January 24, 2026 • Jerry Zhang
Introducing OT-SGN, a novel framework that treats LLM text generation as trajectory optimization on Riemannian manifolds, enabling controlled semantic navigation and cross-domain insight generation.
Read MoreInterstella Project: Research Progress in Semantic Manifold Navigation and Controllable Emergence
January 14, 2026 • Jerry Zhang
A comprehensive review of the Interstella project's research trajectory, from manifold hypothesis validation to geometric navigation frameworks for controllable LLM emergence.
Read MoreWhat is Computable Emergence Engineering?
January 12, 2026 • Jerry Zhang
Computable Emergence Engineering represents Interstella's core vision: transforming AI capability emergence from alchemy into engineered geometric navigation through semantic manifolds.
Read MoreValidating the Semantic Manifold Hypothesis: Token vs Sentence Embeddings
January 12, 2026 • Jerry Zhang
Rigorous analysis of whether LLM semantic spaces satisfy manifold assumptions, with empirical evidence showing sentence-level embeddings are significantly more manifold-like than token-level embedd...
Read MoreProxy Geometric Exploration of Semantic Manifolds in Large Language Models: Evidence for Controllable Emergence Paths
January 12, 2026 • Jerry Zhang
A proof-of-concept framework using proxy manifold learning techniques to explore and control semantic traversal in LLMs, demonstrating controllable 'semantic jumps' via extreme hybrid prompts.
Read More语义流形探索：大型语言模型中可控涌现路径的实证证据
January 12, 2026 • Jerry Zhang
Interstella项目最新研究发布：通过代理几何方法探索LLM语义空间，首次提供可控涌现路径的实证证据，为AGI工程化奠定理论基础。
Read More