nx_face_novelty.nx
buildroot/runtime/nx_face_novelty.nx
about
nx_face_novelty.nx -- MEASURED face-diversity / repetition meter over the
sovereign descriptor store (knowledge/status/galx_face_descr.bin). Answers
the "FLUX-face" question by MEASUREMENT, not guess: are our GENERATED faces
diverse, or is the model emitting the same face repeatedly?
Reads the {cid(72)+16384 i64} records, computes pairwise chi-square distances
(reusing nx_lbp_core chi2_at), and reports the distance distribution + how
many DISTINCT-face clusters survive at a sweep of thresholds. Many near-dup
pairs / few clusters = repetition (flag the recurring synthetic face); wide
spread / many clusters = healthy generation diversity.
IDENTITY-FREE: this only measures self-similarity within OUR OWN content.
No naming, no external lookup. Naming is the operator DATA layer
(galx_models.tsv), attached separately -- never inferred from a face here.
genealogy_id: lbp_chi2_self_similarity + generation_mode_collapse_detection
lineage_id: pairwise_chi2_distribution + threshold_sweep_cluster_count
Build/run: ./_offc/nx_sov_build_run.elf nx_face_novelty license_tier: ORIGINAL
dependencies 2 imports · 0 importers
imports: nx_syscalls.nxnx_lbp_core.nx
imported by: nobody (leaf or entry point)
call flow from main pre-order; caps 40 nodes / depth 6 declared; ↻ = already shown
structs
| none |
consts
| 21 | const K_MAGIC_500000: i64 = 500000 |
| 22 | const K_MAGIC_1000000: i64 = 1000000 |
| 23 | const K_MAGIC_2000000: i64 = 2000000 |
| 24 | const K_MAGIC_4000000: i64 = 4000000 |
| 25 | const K_MAGIC_8000000: i64 = 8000000 |
| 26 | const K_MAGIC_16000000: i64 = 16000000 |
| 28 | const REC: i64 = 131144 |
| 29 | const D: i64 = 16384 |
functions
| 31 | func main() -> i64 |