tractatus-framework/audit-log-viewer

Tractatus Audit Log Viewer

Interactive explorer for 3,942 governance decisions from production deployment of the Tractatus AI Safety Framework.

What is This?

This Space allows you to explore real audit logs from the Tractatus framework - an architectural AI governance system that enforces constraints through structure, not training.

Explore:

• —3,942 governance decisions from production use
• —Six coordinating services (Deep Interlock principle)
• —Context pressure levels (NORMAL → ELEVATED → HIGH → CRITICAL)
• —Tractatus boundary enforcement
• —Service coordination patterns

🎯 Quick Start

Explore the data in 3 ways:

1. 1.By Service - See how BoundaryEnforcer, CrossReferenceValidator, etc. coordinate
2. 2.By Pressure - Filter NORMAL → ELEVATED → HIGH → CRITICAL decisions
3. 3.By Coordination - Find multi-service Deep Interlock examples

🔢 Dataset Statistics

• —Total Decisions: 3,942
• —Date Range: 2024-04 - 2025-11
• —Services Represented: 6 (BoundaryEnforcer, ContextPressureMonitor, CrossReferenceValidator, MetacognitiveVerifier, InstructionPersistenceClassifier, PluralisticDeliberationOrchestrator)
• —Coordination Rate: 1.7% (66 multi-service decisions)
• —Update Frequency: Quarterly (next: 2026-02-01)

Five Architectural Principles

Adapted from Christopher Alexander's pattern languages:

1. 1.Not-Separateness - Governance in critical execution path
2. 2.Deep Interlock - Services coordinate through mutual validation
3. 3.Gradients Not Binary - Intensity-based responses, not yes/no
4. 4.Structure-Preserving - Audit continuity across framework versions
5. 5.Living Process - Evolved from operational failures

Data Privacy

All audit data is anonymized:

• —No file paths
• —No specific code
• —No user information
• —Only governance patterns

📊 Use Cases

For Researchers:

• —Quantitative analysis of governance patterns
• —Service coordination effectiveness studies
• —Context pressure distribution analysis
• —Empirical validation of architectural principles

For AI Practitioners:

• —Real-world governance decision examples
• —Framework design patterns for your own systems
• —Audit logging best practices

For Policy Makers:

• —Evidence of production AI governance
• —Transparency and accountability examples
• —Risk mitigation approaches

Resources

• —Website: https://agenticgovernance.digital
• —Architecture: https://agenticgovernance.digital/architecture.html
• —Documentation: https://agenticgovernance.digital/docs.html
• —GitHub: https://github.com/AgenticGovernance/tractatus-framework
• —License: Apache 2.0

About

Tractatus framework developed in Aotearoa New Zealand. Testing whether architectural principles from the built environment translate meaningfully to AI governance.

Not trying to sell anything - this is active research. Curious what breaks when you poke it.

Build Version: 1.0.1 (2025-10-31)