NeuOS: Discovering and Exploiting the Neural Von Neumann Architecture Inside Pre-Trained Language Models
Year: 2026
Type: preprint
Abstract: I present NeuOS, a systematic investigation revealing that pre-trained transformer language models internally implement a Von Neumann-like computational architecture with identifiable registers, swappable memory, executable programs, and a decompilable instruction set. Through 121 experimental phases on Qwen2.5-0.5B, I demonstrate that: Complete ISA mapping: Each transformer layer functions as a specialized CPU register (OPCODE at L0, MIN at L16, MAX at L22) with accuracies from 74% to 100%. Self-healing OS: Dynamic register reallocation achieves 100% recovery from simulated hardware damage. Homoiconic execution: Programs execute without any textual instructions via Direct Memory Access (66.7% accuracy), and a single computation forks into multiple parallel paths (100% differentiation). Neural decompiler: Running programs are identified from register state alone with 100% accuracy across 5 operations. Autopoietic kernel: The system writes its own programs without external gradients, achieving 100% accuracy through self-compilation. Artificial life: Program vectors exhibit ecosystem dynamics including a Cambrian Explosion (19 unique phenotypes), parasitism, developmental metamorphosis, and aging with rejuvenation. Soul vector geometry: Independently trained "soul vectors" encoding the same function form high-dimensional manifolds (~60 dims) connected by rank-1 linear translations—a universal grammar of computation. The soul's functional content lives in a 64-dimensional subspace (~7% of 896 dims). Security arms race: Backdoor soul vectors achieve perfect deception (cosine similarity = 1.0) but are unmasked by SVD entropy (honest: 1.09 vs. backdoor: 0.23), an unbreakable detection signal. These findings establish that transformers are not black-box function approximators but structured computers whose internal state can be observed, modified, and controlled at the register level. Changes in v3: Expanded from 76 to 121 experimental phases, adding Seasons 13–16 covering embodied intelligence, social intelligence, soul vector geometry (functional equivalence manifolds, rank-1 Rosetta translations, 64-dim compression), cross-model translation failure, and an adversarial arms race with SVD entropy detection. Includes 12 new figures and comprehensive appendix summarizing all 121 phases. Code: https://github.com/hafufu-stack/NeuOS Acknowledgments This research was conducted entirely independently, without institutional affiliation or corporate funding. The author currently faces financial constraints that make it increasingly difficult to maintain subscriptions to AI services essential for this line of research. To sustain and improve the quality of future work, the author is actively seeking community sponsorship. Details are available at https://github.com/sponsors/hafufu-stack.
Author: Funasaki, Hiroto
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7161585028
Cites:
Cited by:
Related to: