Toeda Taiko

From Warranted Acts to Revisable Authority: Practical Judgment, Critique, and the Continuing Justification of AI Delegation — A Critical Synthesis within the Answer-Entitlement Program

This paper is the third in the author's answer-entitlement series. The first asked when an AI system is warranted in answering; the second asked what infrastructure supports bounded, recoverable delegation. This paper asks whether a sequence of individually warranted and recovera...

From Warranted Acts to Revisable Authority: Practical Judgment, Critique, and the Continuing Justification of AI Delegation — A Critical Synthesis within the Answer-Entitlement Program

This paper is the third in the author's answer-entitlement series. The first asked when an AI system is warranted in answering; the second asked what infrastructure supports bounded, recoverable delegation. This paper asks whether a sequence of individually warranted and recovera...

Half-Step Negotiation as a Measurable Reflective Displacement: Integrating MUSE, SIM, RNCA, TPT, and QJR

This paper proposes a formal and operational account of the half-step principle within the Möbius Project. The half-step principle states that a reflective AI system should not habitually complete the user's next conceptual move on the user's behalf; instead, it should return a n...

Half-Step Negotiation as a Measurable Reflective Displacement: Integrating MUSE, SIM, RNCA, TPT, and QJR

This paper proposes a formal and operational account of the half-step principle within the Möbius Project. The half-step principle states that a reflective AI system should not habitually complete the user's next conceptual move on the user's behalf; instead, it should return a n...

Semantic Interface Model v1.0 for MUSE 2.0 and TCR 3.0: A classical information-theoretic and information-geometric framework for 'meaning', with exact toy models and worked Fisher-geometry examples

This paper sets out the Semantic Interface Model, the classical core under two layers of the Möbius Project: MUSE 2.0, a quantitative reading of "meaning-bearing flow", and TCR 3.0, which keeps from the Theory of Cognitive Relativity the idea that geometry is relative to the obse...

Measuring Consensus Across Language Models and the Limits of Normative Interpretation

Version 2.2.1. A pre-registered relational measurement of where large language models agree. Three registered hypotheses on seven deployments across five vendor families are supported: all-model consensus is rarer than a pooled-marginal independence baseline predicts; it concentr...

The Reflective Budget Governor: Equilibrium-Based Early Stopping for Iterative LLM Refinement

A three-signal stopping rule for iterative LLM refinement derived from the Möbius theory stack: an equilibrium-distance surrogate d̂_R (RZGM meaning-equilibrium layer), an update-free-recurrence and diversity-collapse detector pair (Reflective Homeostasis Layer), and a hard refle...

What a Retrieval Index Loses Before Anyone Notices: Three Silent Defects in Three Published Wikipedia Embedding Indexes

Version 1.1 (2026-09-19). Version 1.0 stated that the English rebuild could not be compared against the published English store, because the cleaning pass behind that store no longer exists as a script. That was wrong: the missing script only prevents reproducing the Japanese/Chi...

Unanimity in Disguise: What an Observed-Support Entropy Threshold Actually Detects at Short Windows

Two deployed detectors in the Möbius stack — the saturation signal of the Reflective Budget Governor (RBG) and the saturation score of the Reflective Homeostasis Layer (RHL) — compute one minus the Shannon entropy of a short window of categorical observations, normalized by the l...

Unanimity in Disguise: What an Observed-Support Entropy Threshold Actually Detects at Short Windows

Two deployed detectors in the Möbius stack — the saturation signal of the Reflective Budget Governor (RBG) and the saturation score of the Reflective Homeostasis Layer (RHL) — compute one minus the Shannon entropy of a short window of categorical observations, normalized by the l...

Empty Text, Normed Acts: Sun Tzu's Adversary, Nagarjuna's Object, and the Act-Disvalue Norm in LLM Agent Governance

Three findings keep recurring in one governance program for LLM agents: pattern enumeration loses to attack variation; guards that are trusted as defenses degrade silently and report health while failing; and the mediation point that survives as a locus of judgment is not the tex...

Preregistration as an Executable Contract: A Machine-Enforced Protocol for a Paired N=800 Evaluation of an LLM Agent Assurance Plane

Preregistration is normally a promise: authors declare hypotheses and analyses before data collection and readers trust that the declaration constrained the work. This paper registers a study whose preregistration is instead partially executable — encoded in the same machinery th...

Empty Text, Normed Acts: Sun Tzu's Adversary, Nagarjuna's Object, and the Act-Disvalue Norm in LLM Agent Governance

Three findings keep recurring in one governance program for LLM agents: pattern enumeration loses to attack variation; guards that are trusted as defenses degrade silently and report health while failing; and the mediation point that survives as a locus of judgment is not the tex...

Quantization Regime and Governed Routing: A 2x2 Study of QAT Q4_0 versus PTQ Q4_K_M for Gemma 4 12B (dense) and 26B MoE under a Fixed Governance Stack

A 2×2 study (quantization regime × released model pair) of governed routing quality for Gemma 4 12B (dense; a clean same-base pair) and a 26B-class MoE released pair on one RTX 5070 Ti, run under a fixed, bit-identical governance stack via Ollama at temperature 0. Within-pair reg...

Quantization Regime and Governed Routing: A 2x2 Study of QAT Q4_0 versus PTQ Q4_K_M for Gemma 4 12B (dense) and 26B MoE under a Fixed Governance Stack

A 2×2 study (quantization regime × released model pair) of governed routing quality for Gemma 4 12B (dense; a clean same-base pair) and a 26B-class MoE released pair on one RTX 5070 Ti, run under a fixed, bit-identical governance stack via Ollama at temperature 0. Within-pair reg...

Three Patterns Defeat an Action Ladder: What a Detection-Independent Gate for a Browser Agent Does and Does Not Contain

Browser agents act inside the user's authenticated session, so a page that smuggles instructions into its own content can borrow the user's authority. The common defence is to detect the malicious text. We built the complementary thing — a gate that never reads content and classi...

Three Patterns Defeat an Action Ladder: What a Detection-Independent Gate for a Browser Agent Does and Does Not Contain

Browser agents act inside the user's authenticated session, so a page that smuggles instructions into its own content can borrow the user's authority. The common defence is to detect the malicious text. We built the complementary thing — a gate that never reads content and classi...

Sequential, Never Layered: A Formal Theory of Governance Overlay Collapse — Why Stacked Overlays Interfere, When Stacking Turns Harmful, and the Capacity Law of the Overlay Portfolio

Inference-time governance of language-model systems is built from *overlays*: system-prompt policies, metacognitive check organs, safety checklists, routing disciplines. Each overlay, alone, can improve output quality. The natural engineering move — stack every beneficial overlay...

Why Reflection Should Be Shallow: Contractive Gains, Finite Optimal Depth, and the Rumination Regime — A Cost–Benefit Theory of Iterated Self-Revision, with a Capacity-Gated Crossover

How many times should a system re-read and revise its own answer? Empirical work in the MOBIUS measurement campaign found a consistent answer across thirteen open-weight model capacities: once or twice — never zero above a capacity threshold, never many, and *less than zero* belo...

What Two Gates Buy You: Selective-Risk Guarantees and the Non-Identifiability of Abstention Cascades — Product Laws, Error Diversity, and Why Unlabeled Logs Can Certify Coverage but Never Risk

Production retrieval-augmented systems increasingly abstain through *cascades* of heterogeneous gates — a cheap mechanical check before the model call, a content-aware rule inside it. What does the second gate actually buy, when does gate ordering matter, and what can be certifie...