Independent University of Moscow

The Principle of Maximal Consciousness: From Probabilistic Teleology to the Ultimate Probabilistic Attractor

This paper introduces the Principle of Maximal Consciousness (PMC), a theoretical framework addressing the observer typicality problem in anthropic cosmology. Its central mechanism, the Principle of Proportional Indication (PPI), assigns observer weight according to the total mea...

The Artian Keystone Audit

Which QTT numbers are pure, anchored, keystone-dependent, or structural? The Keystone Audit classifies every headline result before judging its numerical landing. Its paper-native master split remains \[ \boxed{ \mathcal R = \mathcal R_{\rm U}\sqcup \mathcal R_{\rm A}\sqcup \math...

Imagine to Ensure Safety in Hierarchical Reinforcement Learning

This work investigates the safe exploration problem in reinforcement learning, where an agent must maximize cumulative performance while simultaneously satisfying safety constraints. This challenge becomes particularly difficult in long-horizon tasks because safe reinforcement le...

rmt-llm-research — Random Matrix Theory meets Large Language Models: spectral analysis of activations and the mathematical inevitability of hallucinations

A research repository by Iskhak Hamzatovich Isaev presenting a comprehensive program applying Random Matrix Theory (RMT) to the analysis of large language models (LLMs). Thread 1 — Spectral statistics of LLM activations. Covariance matrices of GPT-2 hidden-state activations exhib...

rmt-llm-research — Random Matrix Theory meets Large Language Models: spectral analysis of activations and the mathematical inevitability of hallucinations

A research repository by Iskhak Hamzatovich Isaev presenting a comprehensive program applying Random Matrix Theory (RMT) to the analysis of large language models (LLMs). Thread 1 — Spectral statistics of LLM activations. Covariance matrices of GPT-2 hidden-state activations exhib...