Tony Mason

Tony Mason

System prompt instructions that cooperate in English compete in Spanish, with the same semantic content, but opposite interaction topology. We present instruction-level ablation experiments across four languages and four models showing that this topology inversion is mediated by social register: the imperative mood carries different obligatory force across speech communities, and models trained on multilingual data have learned these conventions. Declarative rewriting of a single instruction block reduces cross-linguistic variance by 81% (p = 0.029, permutation test). Rewriting three of eleven imperative blocks shifts Spanish instruction topology from competitive to cooperative, with spillover effects on unrewritten blocks. These findings suggest that models process instructions as social acts, not technical specifications: "NEVER do X" is an exercise of authority whose force is language-dependent, while "X: disabled" is a factual description that transfers across languages. If register mediates instruction-following at inference time, it plausibly does so during training. We state this as a testable prediction: constitutional AI principles authored in imperative mood may create language-dependent alignment. Corpus: 22 hand-authored probes against a production system prompt decomposed into 56 blocks.

Tony Mason

We find that models report highest confidence precisely when they are fabricating. Across four model families (OLMo-3, Llama-3.1, Qwen3, Mistral), self-reported confidence inversely correlates with accuracy, with AUC ranging from 0.28 to 0.36 where 0.5 is random guessing. We prove, under explicit formal assumptions, that this is not a capability gap but an observational one. Under text-only observation, where a supervisor sees only the model's output text, no monitoring system can reliably distinguish honest model outputs from plausible fabrications. We prove two results: first, that any policy conditioning only on the query cannot satisfy epistemic honesty across ambiguous world states; second, that no learning algorithm optimizing reward from a text-only supervisor can converge to honest behavior when the supervisor's observations are identical for both grounded and fabricated responses. Within our formal model, these impossibilities hold regardless of model scale or training procedure, including RLHF and instruction tuning. We construct a tensor interface that escapes the impossibility by exporting computational byproducts (per-token entropy and log-probability distributions) that are structurally coupled to correctness under standard training. Per-token entropy achieves pooled AUC 0.757, outperforming all text baselines by 2.5--3.9 percentage points at every budget level tested (10\%, 20\%, 30\%). The entropy signal generalizes across architectures (Spearman $ρ= 0.762$). The core contribution is a cost surface where the empirical mapping from verification budget (fraction of queries receiving expensive checks) to detection accuracy for each judge strategy is a practical lookup for system builders deciding how to allocate verification resources. The contribution is the map. The territory is the system you are building.

Tony Mason

Quorum design over asymmetric topologies conflates two independent concerns: inter-tier obligation (which tiers must participate for cross-tier safety) and intra-tier replication (how each tier survives local failures). Flat quorums treat all nodes as interchangeable; when consensus fails, the structure does not reveal whether a tier was unreachable or a tier lost too many replicas. We show that mapping a crumbling-wall quorum construction to a physically tiered network separates these concerns and makes the protocol's failure modes legible: an operator can determine which tiers retain global consensus capability from the wall structure and connectivity state alone, without runtime probing. Using a 10-node Earth/LEO/Moon/Mars topology as a magnifying glass, we confirm that three of four tiers retain global liveness during Mars conjunction blackout; only the disconnected tier loses it. Consensus latency at each tier equals the speed-of-light round-trip to Earth: 183~ms (Earth), 131~ms (LEO), 5.1~s (Moon). The wall also imposes a leadership cost gradient on Multi-Paxos elections that symmetric grid quorums cannot express. A comparison between sparse and full-coverage topologies separates wall obligations from network reachability as independent liveness constraints. All results are design-level; quorum intersection is verified exhaustively in TLA+.

Tony Mason

Leyva-Vázquez and Smarandache (2025) demonstrated that neutrosophic T/I/F evaluation, where Truth, Indeterminacy, and Falsity are independent dimensions not constrained to sum to 1.0, which reveals "hyper-truth"' (T+I+F > 1.0) in 35% of complex epistemic cases evaluated by LLMs. We extend their work in two directions. First, we replicate and extend their experiment across five model families from five vendors (Anthropic, Meta, DeepSeek, Alibaba, Mistral), finding hyper-truth in 84% of unconstrained evaluations, which confirms the phenomenon is cross-vendor under our prompt protocol. Second, and more significantly, we identify a limitation of scalar T/I/F that their framework cannot address: models adopting an `"Absorption" position (T=0, I=1, F=0) produce identical scalar outputs for fundamentally different epistemic situations (paradox, ignorance, contingency), collapsing the very distinctions neutrosophic logic was designed to preserve. We demonstrate that extending the evaluation to include declared losses (structured descriptions of what the model cannot evaluate and why) substantially recovers these distinctions. Models producing identical scalars for paradox and ignorance produce nearly disjoint loss vocabularies (Jaccard similarity < 0.10 on loss description keywords), with domain-specific, severity-rated loss declarations that differentiate the nature of their uncertainty. This suggests that scalar T/I/F is a necessary but insufficient representation of epistemic state, and that tensor-structured output (scalars + losses) provides a more faithful model of LLM epistemic capabilities.