What question did this study set out to answer?

To create a governance layer that ensures AI outputs are stoppable and non-drifting, particularly in the context of missing evidence.

February 19, 2026Open Access

FCL-S Inference Kernel (MVP+): Inference-Time Stop/Attribution Governance for Non-Commitment under Missing Evidence (UST/SCB/AF) Inference-time governance for stoppable, non-drifting outputs

Key Points

To create a governance layer that ensures AI outputs are stoppable and non-drifting, particularly in the context of missing evidence.
Introduced the FCL-S Inference Kernel to govern AI decision-making outcomes.
Implemented core components like UST, SCB, and AF to control AI output addressing coherence and errors.
Developed MVP+ for detecting confidence mismatches and semantic drift using defined thresholds.
Established an evaluation protocol for assessing governance effectiveness against various error metrics.
Successfully constrained AI output under conditions with missing evidence.
Demonstrated improvements in preventing authority bias and semantic drift.
Quantified confidence mismatch under specified thresholds, reducing false positives in output contradictions.

Abstract

This deposit provides an implementation-spec MVP+ reference for the FCL-S (False-Correction Loop Stabilizer) Inference Kernel, a minimal inference-time governance layer that makes advanced AI stoppable and non-driftingbefore scaling capability. Problem addressed. In the post-scaling regime, systems can produce highly plausible coherence that locks in errorsand amplifies misinformation through interaction-level self-correction. This deposit specifies a wrapper-style kernel that constrains when a generator may emit a committing answer. Core components (UST/SCB/AF). UST (Unknown Stable Terminal): “Unknown” as a governed terminal state. Returns structured missing slotsand safe next steps for verification. SCB (Stop/Correction Boundary): a stop boundary evaluated in pre and post phases (dual-wall envelope) that blocks ungrounded commitment, drift, or contradictions and routes to UST. AF (Attribution Fixation): structurally separates primary definitions (author-defined) from external observations/inferences, preventing authority-bias drift. MVP+ extensions. Confidence mismatch detection using token-confidence proxies (e. g. , avgₜokenₗogprob) under a heuristic threshold (τ = −2. 5). Semantic drift scoring (driftₛcore) against primary definitions (threshold 0. 35). Unresolved-UST gating (openᵤst) to reduce false-positive contradiction triggers. MVP++ specification notes: quiet-drift detection via embedding stack, and governance strength control via strictₘode. Evaluation protocol (summary). This deposit includes a compact test plan and metrics covering ungrounded commitment, confidence mismatch, authority drift, semantic drift, long-context failure, and cumulative agent errors.

Bookmark

View Full Paper

Bookmark

View Full Paper

FCL-S Inference Kernel (MVP+): Inference-Time Stop/Attribution Governance for Non-Commitment under Missing Evidence (UST/SCB/AF) Inference-time governance for stoppable, non-drifting outputs

Key Points

Abstract

Cite This Study