论文集

ORIGINAL THOUGHT PAPER · JULY 2026 · V3 AI安全方向是什么？ From Alignment Tax to Incremental Information Productivity:A Paradigm Rethinking of AI Safety Direc…

ORIGINAL THOUGHT PAPER · JULY 2026 · V2 从智力到智能 From Intellect to Intelligence: A Human Higher-Order Cognitive Transformation Model Based on Incrementa…

POSITION PAPER · JULY 2026 超越认知复杂度迈向认知创造的热力学理论 以 Epiplexity 为起点，构建负熵认知体的信息论框架 Beyond Epiplexity: Toward a Thermodynamics of Cognitive CreationBuilding…

CRITICAL ANALYSIS REPORT · JULY 2026 · V1 Anthropic！新AI共产党化巨头 安全过滤器的精度危机、用户侮辱机制与地域歧视代码的实证研究 Anthropic: The New AI Authoritarian GiantAn Empirical Stud…

ORIGINAL THOUGHT PAPER · JUNE 2026 金融投资作为量产债务奴仆的工具 从信用杠杆叠加到系统性财富转移的机制研究基于1637–2026年金融危机的跨历史分析 Financial Investment as a Mass-Production Tool for Debt …

ORIGINAL THOUGHT PAPER · JUNE 2026 · V4 Transformer 架构的本体论矛盾 从 Softmax 零和博弈到物理世界映射断裂一个贯穿 LLM 已知问题的统一信息论框架 The Ontological Contradiction of Transformer…

ORIGINAL THOUGHT PAPER · JUNE 2026 · V1 AI的Output与Input结构性元认知级别分析 基于逆向工程方法的ChatGPT 5.5 Thinking High四段对话行为深度解构 Structural Meta-Cognitive Analysis of A…

ORIGINAL THOUGHT PAPER · JUNE 2026 · V4 异质Agent是范式性革命 Heterogeneous Agents as a Paradigmatic Revolution:From Model-Centrism to Orchestration-Centrism …

ORIGINAL THOUGHT PAPER · JUNE 2026 · V3 AI是同质单一系统 AI is a Homogeneous Monolithic System: Architectural Uniformityas a Key Contributing Factor in Epist…

ORIGINAL THOUGHT PAPER · JUNE 2026 · V2 ChatGPT的抽象层多个问题分析报告 基于逆向工程方法的ChatGPT 5.5 Thinking High对话行为深度解构 Reverse-Engineering ChatGPT’s Abstraction…

ORIGINAL THOUGHT PAPER · JUNE 2026 · V2 有限认知与无限现实 从统计压缩的有限表征不可内生恢复原理到人类元认知的存在论边界——统一统计学、人工智能与人类认知有效性边界的理论框架 Finite Cognition and Infinite Reality:From…

MODEL FIELD REPORT · JUNE 2026 · V3 Fable 5 行为指纹实测报告 Fable 5 Behavioral Fingerprint Report: From Zero-Day Vulnerability Searchto Formal-Logic Vulnerab…