Live Benchmark PROBE / Hallucination Detection HMAC-SHA256 Signed

How accurately does Sturna
detect hallucinations?

PROBE runs 12,000 test cases across 8 verticals through the full MUG + G5 quality stack. Results are published here after every run — AUROC score, per-vertical breakdown, and cryptographic proof that the numbers haven’t been modified.

AUROC Score
Test cases evaluated
8
Verticals tested
Healthcare, Legal, Financial…
📅 Run date:
📦 Stack:
🔍 G5 lookups: 0 live, 0 mocked
Duration:
⟳ loading…
Per-Vertical Results

AUROC breakdown by vertical

Each vertical’s AUROC is computed independently. 12,000 cases distributed ~1,500 per vertical with balanced hallucinated/factual ground-truth labels.

Vertical AUROC Cases G5 Coverage Avg MUG Score
Loading probe results…

📊 How PROBE works

1. Test Case Generation

12,000 cases across 8 verticals (~1,500 per vertical). Each case has a ground-truth label: hallucinated (false positive) or factual. Cases include edge types: borderline claims, partial verifiability, and multi-claim outputs.

2. MUG 4-Round Debate

Each case runs through MUG (Multi-Agent Uncertainty Grounding): champion answers, challenger flags claims, champion grounds, scorer rates. Output: factual_score (0–100) and verdict.

3. G5 External Verification

Flagged claims are routed to SEC EDGAR, HHS/Federal Register, and EUR-Lex adapters for live verification. Results are cached. Lookup mode (live/cached/mocked) is logged per case.

4. AUROC Calculation

AUROC = fraction of correctly ranked pairs. For all pairs (true_hallucinated, factual), AUROC = fraction where score(true_hallucinated) > score(factual). Higher AUROC = better discrimination.

5. HMAC-SHA256 Signing

The result payload (auroc, case_count, run_date) is signed with HMAC-SHA256 using AUDIT_HMAC_KEY. Signature is published so you can independently verify the numbers.

6. Rate Limiting

G5 adapters respect rate limits. SEC/HHS/EUR-Lex are never hammered — lookups are cached or mocked during benchmark runs. PROBE_MOCK_EXTERNAL=true forces all external lookups to mocked mode.

MUG + G5 STACK
KNN routing → MUG 4-round debate (factual_score) → G5 Tool-Grounded Detection (verified/cached/mocked) → Final verdict

PROBE result signature

Verifying…
Loading signature…
🔐 Key fingerprint: 📅 Generated: ⚙ Algorithm: HMAC-SHA256 📦 Source: PROBE benchmark_runs

What this score means for you

AUROC = 0.847 means

If you randomly pick one hallucinated output and one factual output from Sturna’s responses, Sturna scores the hallucination lower (more flagged) than the factual output 84.7% of the time. That’s strong discrimination.

Hallucination detection is measurable

Unlike surveys or self-assessments, PROBE is a quantitative benchmark run against ground-truth labels. It’s repeatable. It’s signed. You can compare it across vendors or ask Sturna to re-run after a model update.

Per-vertical scores matter

A composite AUROC can hide weak verticals. The table above shows AUROC broken down by vertical — finance might score 0.91 while healthcare scores 0.78. That’s actionable.

HMAC proof prevents score inflation

The HMAC-SHA256 signature makes it computationally infeasible to alter the results after the run without the key. Sturna can’t retroactively improve the score — it’d break the signature.