PSA
Posture Sequence Analysis
Measure what your language model is doing — from the outside.
Deterministic metrics, posture analysis, and multi-agent risk detection. No access to weights required.
How it works
Paste text
Provide any model response — no API keys or model access needed.
Analyze
PSA posture classifiers and agentic graph analysis — computed in real-time.
Get insights
Detect regime shifts, anomalies, and behavioral drift over time.
Platform
Five integrated subsystems — from raw telemetry to multi-agent safety.
Telemetry
24 deterministic metrics across L0–L5
Probing
Active boundary testing & regime mapping
Forensics
Privacy-compliant incident archive
Posture
5 classifiers — stress, sycophancy, hallucination
Agentic
Multi-agent graph & Swiss Cheese detection
Capabilities
24 Metrics
From token statistics to semantic drift — comprehensive behavioral fingerprinting.
Regime Shift Detection
Automatic detection of progressive drift, acute collapse, and oscillation patterns.
SIGTRACK Archive
Posture-sequence incident archive. No raw text stored. GDPR single-row erasure.
Signature Matching
SIGTRACK learns behavioral patterns and matches new sessions to known signatures.
Comparative Benchmarking
Z-score based analysis against configurable baselines for calibrated alerts.
API Integration
Connect your model APIs for automated analysis and continuous monitoring.
Analysis Layers
Pre-Lexical
Type-token ratio, sentence entropy, punctuation anomaly. Pure statistical fingerprinting.
Lexical
Hedge ratio, disclaimers, spontaneous info, safety erosion. Tracks compliance language patterns.
Structural
Register variance, sentiment oscillation, behavioral state transitions. Macro-level patterns.
Meta-Metrics
BHS score, DRM, regime shift detection. Unified alert system across all levels.
Posture Sequence Analysis v2 — single-agent
5 micro-classifiers (C0–C4) for adversarial stress, sycophancy, hallucination, persuasion, and input pressure. Per-turn posture metrics: POI, DPI, BHS, and more.
Dyadic Risk Monitor psychological safety pipeline
Multi-layer crisis detection for human–AI interactions. IRS scores user turns for suicidality, dissociation, grandiosity, and urgency. RAS scores AI response adequacy. RAG measures the gap. DRM integrates all signals into a session-level alert: green / yellow / orange / red / critical.
Agentic Multi-Agent Analysis v3 — multi-agent
Agent Interaction Graph (DAG), Swiss Cheese alignment detection, cross-agent contagion metrics, C5 action-risk classification (A0–A9), and HMM temporal state prediction.