Historical task source. The Eve task board holds current task state.
Phase 90: ML Sovereignty — Alignment & Safety Training#
Build safety guardrails and alignment training pipelines to ensure all self-hosted models meet Oshun's safety standards.
90.1 Safety Models & Guardrails#
90.1.1 Safety Infrastructure#
- 90.1.1.1 Integrate Llama Guard 4 as primary guardrail model (12B, natively multimodal text+image, MLCommons hazards taxonomy, single GPU)
- 90.1.1.2 Integrate Granite Guardian for RAG hallucination detection (0.854 AUC — SOTA hallucination, 0.871 risk detection)
- 90.1.1.3 Build Constitutional AI training pipeline with Oshun-specific principles (safety, helpfulness, honesty, privacy, fairness — per domain customization)
- 90.1.1.4 Build red-team automation pipeline (auto-generate adversarial prompts → test model → classify failures → add to safety training → retrain → repeat)
- 90.1.1.5 Implement RLAIF pipeline using Oshun's own models as judges (eliminate external API dependency for preference labeling)
- 90.1.1.6 Build jailbreak detection fine-tuning pipeline (collect attempts from production + synthetic → train detection model → deploy as input guardrail)
- 90.1.1.7 Build prompt injection detection model (fine-tune classifier on injection attempts from Kuan Yin + synthetic generation)
- 90.1.1.8 Implement safety benchmark continuous evaluation (TruthfulQA, BBQ, RealToxicityPrompts — run on every model version)
- 90.1.1.9 Build deepfake detection pipeline (integrate Deepfake-o-Meter framework with 18+ detection models across image/video/audio)
- 90.1.1.10 Build NSFW content classifier fine-tuning pipeline for Aphrodite domain (domain-specific boundary calibration)
- 90.1.1.11 Implement NeMo Guardrails integration (Content Safety NIM + Topic Control NIM + Jailbreak Detection NIM)
- 90.1.1.12 Build multi-lingual safety evaluation for Ghanaian languages (safety benchmarks adapted for Twi, Ewe, Ga, Hausa)
90.1.2 Content Provenance, Watermarking & IP Compliance#
- 90.1.2.1 Build text watermarking system for LLM outputs (statistical watermark embedding during token generation: green/red list method, robust to paraphrasing, detectable without model access, configurable strength, per-model watermark keys)
- 90.1.2.2 Build image/audio watermarking for generated media (invisible watermark embedding in Isis-generated images and Euterpe audio, robust to compression/cropping/transcoding, detection API endpoint)
- 90.1.2.3 Integrate C2PA Content Credentials for all generated content (embed provenance manifests: creation tool, AI model used, generation parameters, timestamp — for images, audio, video from Isis/Euterpe/Psyche, EU AI Act Article 50 compliance)
- 90.1.2.4 Build copyright compliance system for generated content (training data license tracking from 87.2.1.5 → output license attribution, near-duplicate detection against copyrighted works, opt-out compliance)
- 90.1.2.5 Build IP risk scoring for generated outputs (similarity scoring against known copyrighted works per domain, automated flagging above configurable threshold, per-domain IP policy enforcement)
90.1.3 Domain-Specific Safety & Output Validation#
- 90.1.3.1 Build per-domain safety policy framework (domain-specific risk taxonomies and enforcement rules: medical with clinical safety, financial with regulatory compliance, creative with content boundaries — policy-grounded guardrails)
- 90.1.3.2 Implement domain-aware guardrail routing (route safety checks through domain-specific policy evaluators based on request context: Hathor creative vs Sophia research vs Aphrodite adult — different thresholds per domain)
- 90.1.3.3 Build output validation and compliance layer (schema conformance for structured outputs, format validation for domain-specific outputs, factual grounding verification for RAG responses, citation accuracy checking, hallucination detection beyond Granite Guardian)
Completion evidence#
All 20 locally actionable items are implemented in
libs/nous/training/src/phase-90-*.ts. The capability catalog maintains an
exact one-to-one mapping from every checklist ID to implementation and automated
verification. Tests cover model bindings, sovereign training loops, governance
gates, watermarking and C2PA delegation, IP policy, domain routing, structured
output validation, grounding, and citation verification.
Verification: nx test nous-training, nx lint nous-training,
nx build nous-training, and nx typecheck nous-training.