# Эксперимент: Single-Node Anti-Oscillation Anchor (#829)

{
  "worked": false,
  "confidence": 10,
  "verdict": "Built a 16-node weighted bipartite linear recurrent-GNN toy with mode-selected and random single-coordinate anchors. The eigenvalue sensitivity prediction was validated: predicted slope -0.2967 versus observed -0.2993 (0.9% error). However, for the unstable flip mode, the prescribed update J-kappa E_a moves the multiplier farther below -1: spectral radius increased from 1.080 at kappa=0 to 1.451 at kappa=0.8, with no stable gain found; therefore the proposed anti-oscillation mechanism did not manifest.",
  "metrics": {
    "baseline": "No anchor: rho=1.0800, A_K=1.3343, norm_K=9.3531 at 45 steps.",
    "idea": "Mode-selected anchor at kappa=0.2: rho=1.1480, A_K=5.1964, norm_K=21.5607; random anchor: rho=1.0920, A_K=0.8980, norm_K=6.85598. Selected anchoring worsened the unstable flip dynamics."
  },
  "how_to_run": "/home/maxwelhelp/main/bin/python3 anchor_experiment.py",
  "files": [
    "anchor_experiment.py",
    "results.json"
  ],
  "limitations": "This was a linear toy rather than a trained nonlinear recurrent GNN, and it did not test the sign-corrected controller h_next=F(h)-kappa E_a(F(h)-h_ref) or positive feedback for a negative discrete-time mode. No task accuracy or GPU timing comparison was performed."
}