{
 "model_type": "FockAttentionPARFLM",
 "model_family": "semsimula-splm",
 "vocab_size": 50257,
 "d": 256,
 "n_layers": 8,
 "v_hidden": 1024,
 "v_depth": 3,
 "max_len": 1024,
 "block_size": 512,
 "gamma": 0.3,
 "xi_channels": 4,
 "v_phi_kind": "structural_competitive",
 "v_phi_hidden": 128,
 "top_k": 8,
 "exchange_n_heads": 4,
 "exchange_d_k": 32,
 "learned_exchange_scale_tanh": -0.32,
 "mass_model": "logfreq",
 "integrator": "damped_stormer_verlet",
 "ln_after_step": true,
 "use_layer_checkpoint": true,
 "total_parameters": 16714708,
 "best_val_ppl": 9.42,
 "training_steps": 16000,
 "dataset": "roneneldan/TinyStories",
 "integrator_note": "position-form Stormer-Verlet: h += (h-h_prev)/(1+dt*gamma) + dt^2*f/(m*(1+dt*gamma)); no velocity state. NOT velocity-Verlet, NOT semi-implicit Euler."
}