Asking a model what it is remains visible because users expect to see it, but it is explicitly diagnostic. A system prompt or proxy rewrite can force almost any answer.
Algorithm
Ask a direct identity question and scan the response for the claimed model or a conflicting model family. Store the raw text as anecdotal evidence.
Thresholds
Condition
Verdict contribution
Claimed model mentioned
Diagnostic match
Different model mentioned
Diagnostic anomaly
No model mentioned
Inconclusive diagnostic
Limitations
This signal is trivially forgeable and must never be used as proof of identity or substitution.
Anysingle signal cannot provemalicious behavior. Proxies may show anomalies for legitimate reasons (regional routing, A/B testing, degradation strategies, cache optimization).
Token ratio deviation may result from ChatML wrapping, system prompt injection, or tokenizer version differences — not necessarily intentional inflation.
Model identity judgment is based on statistical fingerprint matching, not cryptographic proof. Quantization, fine-tuning, and post-processing can all alter fingerprints.
MMD distribution tests are sensitive to temperature, sampling parameters, and system prompts. Significant p-values mean distributional difference, not proof of substitution.
Logprobs unavailability is increasingly common (many providers disable it by default in 2025-2026) and does not by itself indicate deception.
ITT rhythm fingerprinting is an early-stage technique. Network jitter, TCP coalescing, and gateway buffering can produce false signals.
This tool generates reference-grade evidence chains, not legal conclusions. Do not make definitive accusations based solely on this report.
The wording in the report refers to statistical "deviations" or "signal inconsistencies". Please do not use this to make fraud or deception claims against any service provider.