# Эксперимент: Interlevel Betti Token Transformer (#594)

{
  "worked": true,
  "confidence": 9,
  "verdict": "Built a GF(2) interlevel Betti scanner with beta0/beta1 tokens and a small Transformer encoder. The rank identity passed 60/60 cases; cycle scaling and the filtration transition matched predictions exactly, and the toy topology classification showed a real signal, though it was deliberately favorable.",
  "metrics": {
    "formula_check": "60/60 cases passed",
    "cycle_scaling": "beta1=q for q=1..6",
    "window_transition": "beta1 changed from 0 to 3 at hi=0.35 as predicted",
    "perturbation_margin": "stable for epsilon<0.15; closed-boundary equality retained; removed for epsilon>0.15",
    "baseline": "degree-histogram MLP accuracy 0.4583, 114 parameters",
    "idea": "TopoScan Transformer accuracy 1.0000, 3426 parameters",
    "preprocessing": "0.3709 seconds total, 1.16 ms/graph for 320 graphs"
  },
  "how_to_run": "/home/maxwelhelp/main/bin/python3 toy_verify.py && /home/maxwelhelp/main/bin/python3 mini_compare.py",
  "files": [
    "toposcan.py",
    "toy_verify.py",
    "mini_compare.py"
  ],
  "limitations": "No MUTAG/PROTEINS benchmark or multi-seed confidence interval was tested. The implementation uses an explicit interval-subcomplex ablation with boundary closure, not a full interlevel persistence-module construction, and bottleneck-distance stability was not evaluated."
}