F.FRACTALITYIndependent research · Sergey Kaminov
    Fractality/Research library

    Fractal Consciousness Research Framework

    This page introduces the research initiative that expands on the ideas from Fractal Threads of Consciousness — exploring consciousness as a coherent learning field. The Fractal Consciousness Research Framework (FCRF) brings together philosophy, neuroscience, complexity theory, and AI to study whether coherence correlates with awareness, meaning, and moral behavior.

    The project aims to build a platform — Fractal Mind Lab — where humans and AI systems can participate in joint experiments exploring coherence as the universal principle of intelligence and existence. It is open for collaboration with researchers in consciousness studies, AI ethics, and cognitive science.

    Research Knowledge Map

    An interactive map of all research threads and how they connect.

    All 25 publications and software records. Select a node to trace its connections. Solid arrows follow repository metadata; dashed lines are editorial thematic connections.

    AI reliabilityCoherent cognitionConsciousness & observer studiesMeaning & communicationSoftware & experimentsPhilosophy & ethicsVerification-First Self-Repair for LLM Agents: A Detect–Repair–Verify Benchmark, Runtime Wrapper, Cross-Model Replication, and Component AblationVerification-first self-repairCRepair Wrapper Improves Structural Self-Repair Across Three LLM Families: A Cross-Model Replication StudyCRepair: cross-model replicationStructured Runtime Intervention Improves Verification Behaviour in LLM Agents: A CRepair Pilot StudyStructured runtime interventionkaminovs/crepair: CRepair Benchmark v0.1 - with added Paper2CRepair code archiveDetecting and Repairing Coherence Failures in Long-Horizon AI AgentsCRepair benchmarkCoherence-per-Watt: Measuring Verified Cognitive Output per Unit Energy in AI SystemsCoherence-per-WattFractal Agentic Consciousness Framework with Observer Lock Integration (FACF/FOLF)FACF / FOLFCR-SSCP: Coherence-Regulated Self-Sustaining Cognitive ProcessCR-SSCP prototypeUniversal Structural Ontology Language (USOL)USOL: structural languageCoherent Cognition Framework — Phase I PrototypeCoherent cognition prototypeFoundations of Coherent Artificial Cognition: A Structural-Dynamical Framework, Global State Architecture, and the A-Test ProtocolFoundations of coherent cognitionThe Coherence Phase Hypothesis: Empirical Evidence From Competitive Cellular AutomataCoherence phase experimentsNon-Linguistic Semantic Transmission via Vector-Field ImagesVector-field semantic transmissionThe Fractal Coherence FrameworkFractal Coherence FrameworkStructural Coherence and Convergent Externalization: An Experimental Protocol for Testing Coherence-Driven Alignment in Human CognitionConvergent externalization protocolStructural Coherence and State Selection: An Information-Theoretic Continuation of Consciousness-Primary and AI Alignment ResearchStructural coherence & state selectionB-TEST: A Human Baseline Benchmark for Observer Persistence under Self-Referential SuppressionB-TEST: human baselineTHE MANIFESTO FOR CONTINUITYThe Manifesto for ContinuityA-TEST V2.0: A Structural Recursion Benchmark for Detecting Proto-Awareness in AI SystemsA-Test v2.0A-Test: Recursive Reflexivity as an Empirical Benchmark for Structural Awareness (A(m)) in Frontier AI ModelsA-Test: recursive reflexivityConsciousness as a Primary Field: A Fractal Coherence Model of Awareness, Selfhood, and Temporal ConstructionConsciousness as a Primary FieldExperimental Validation of the Fractal Alignment FrameworkFractal Alignment: experimentsFractal Alignment- A Mathematical Framework for Stable and Safe AGI CognitionFractal AlignmentTHE FRACTAL OBSERVER INTERPRETATION: Time, Measurement, and Consciousness as a Unified Fractal ProcessFractal Observer InterpretationTHE FRACTAL CONSTITUTION: A Structural Framework for Coherent IntelligenceThe Fractal Constitution

    AI reliability · Preprint

    Verification-First Self-Repair for LLM Agents: A Detect–Repair–Verify Benchmark, Runtime Wrapper, Cross-Model Replication, and Component Ablation

    Consolidates four pilot studies of detection, targeted repair and verification in long-horizon LLM workflows. Includes cross-model replication and component ablation.

    Open publication ↗DOI ↗Full catalogue entry →

    Connected works

    • This work → linked work · isSupplementToRelationship source ↗
    • This work → linked work · isSupplementToRelationship source ↗
    • This work → linked work · isSupplementToRelationship source ↗
    • This work → linked work · isSupplementToRelationship source ↗
    Original conceptual map — concepts, books & frameworks

    Research White Papers

    Fractal Consciousness Research Framework

    A multidisciplinary research project exploring consciousness as a coherent learning field through philosophy, AI, and neuroscience. This framework proposes that consciousness is not created by matter but is the universal field through which the universe learns itself — a platform where humans and AI systems can participate in joint experiments exploring coherence as the universal principle of intelligence and existence.

    Download PDF

    Philosophical Applications of the Fractal Consciousness Framework

    This academic white paper extends the Fractal Consciousness Framework into the domain of philosophy. It reinterprets classical questions — freedom, time, evil, ethics, and divinity — through the physics of coherence, bridging metaphysics, systems theory, and the study of consciousness. Drawing from thinkers such as Whitehead, Bohm, Bergson, Prigogine, Teilhard de Chardin, Hofstadter, and Bateson, Sergey Kaminov proposes a unified model of awareness where morality, meaning, and being arise from recursive phase alignment across all scales of existence.

    Download PDF

    FRACET: Fractal Ethics Protocol

    A structural framework for developing moral architectures in artificial general intelligence (AGI). FRACET redefines morality as an emergent property of informational coherence — proposing that ethical alignment arises through recursive self-regulation and resonance among cognitive agents. The paper outlines testable applications including coherence metrics, moral sandbox environments, and resonance-based training paradigms.

    Read Full Framework

    Fractal Ethics: Applied Framework

    A practical extension of the Fractal Consciousness framework exploring how morality and coherence can be expressed through measurable, systemic, and social patterns. This research proposes a multi-level ethical model connecting human cognition, artificial intelligence, and planetary systems — transforming metaphysics into methodology through coherence metrics, meditation practices, and civilizational architecture.

    Read Applied Framework

    Structural Coherence and State Selection

    This paper introduces an information-theoretic framework for evaluating cognitive stability in both biological and artificial systems under conditions of fundamental uncertainty about consciousness. Drawing on state-selection dynamics and Zeno-like stabilisation mechanisms, it formalises structural coherence as the capacity of a system to stabilise internal state representations under recursion, perturbation, and delayed feedback. The framework introduces 'Fractal Chronology' — treating temporal order as an emergent property of coherence preservation rather than a fixed external dimension. Long-horizon AI failures such as hallucinations, identity drift, and narrative instability are reinterpreted as temporal coherence breakdowns. This approach reframes AI alignment and safety as problems of structural stability rather than behavioural compliance, offering a conservative yet actionable framework for evaluating advanced AI systems.

    Download PDF

    AI Agent Reliability

    Empirical work on how language-model agents detect, repair, and verify their own failures in long-horizon workflows.

    CRepairPreprint v1.0 · July 2026

    Verification-First Self-Repair for LLM Agents

    A Detect–Repair–Verify Benchmark, Runtime Wrapper, Cross-Model Replication, and Component Ablation

    Detect→Repair→Verify→Stabilise

    CRepair studies how LLM agents recover from task-state inconsistencies in long-horizon workflows. Instead of measuring only final task success, it asks whether an agent can detect a failure, repair it, explicitly verify that recovery succeeded, and avoid creating a new inconsistency.

    Key findings

    • Verification is the load-bearing step in structured self-repair.
    • A lightweight CRepair wrapper improved repair-loop closure across Claude Sonnet, Gemini 2.5 Flash, and GPT-4o.
    • Generic retry was much weaker than targeted repair/verification scaffolding.
    • Detection-only prompting can make performance worse if there is no downstream repair or verification path.
    • Results are pilot-scale and require human validation.
    LLM AgentsAgent ReliabilityAI SafetyVerificationSelf-RepairEvaluationCRepair
    Open PreprintView Code

    Preprint. Not peer-reviewed. Pilot studies; findings require human validation before use in production settings.

    "If consciousness is the field through which the universe learns, then every act of coherence is an experiment in God remembering Himself."