Empirical Snapshot
CogniPrint currently has a working empirical evidence package supporting a follow-up manuscript.
5controlled perturbation campaigns
41comparison rows in the current package
11campaign-004 comparison rows
Released Benchmark Subset
The first released public benchmark subset remains intentionally small and should be read as a benchmark-growth layer rather than as a broad validation endpoint.
6released benchmark baselines
36released benchmark variants
3released languages in the v1 subset
Validation v1.1
The current statistical validation layer is descriptive. It helps review benchmark-versus-campaign behavior, but it does not by itself justify stronger inferential framing.
9benchmark baselines in validation v1.1
54benchmark variants in validation v1.1
6shared benchmark–campaign bridge axes
Interpretation Guardrail
The evidence visibility layer is descriptive and review-oriented. It does not upgrade the scientific meaning of the current public evidence package into a settled empirical endpoint, a forensic workflow, or a stronger attribution claim.