1experiment:2 name: labels_103m3 seed: 14 mode: prisme5 model: from_checkpoint6 checkpoint: experiments/runs/editor_pilot_103m_lr3e4_seed01_20261030021200/checkpoint7 use_policy: false8 data:9 kind: tokens10 collect:11 batch_size: 812 seq_len: 51213 max_prefix: 314 audit_every: 415 partitions:16 train:17 bin: data/processed/fineweb_edu_300m/train.bin18 n_batches: 12019 stream_seed: 77720 calib:21 bin: data/processed/fineweb_edu_300m/val.bin22 n_batches: 3023 stream_seed: 77824model:25 name: prisme_103m26 vocab_size: 3276827 d_model: 76828 n_layers: 1229 n_heads: 1230 n_kv_heads: 331 ffn_width: 204832 max_seq_len: 409633 rope_theta: 10000.034 norm_eps: 1.0e-0535 tie_embeddings: true36 init_std: 0.0237 use_workspace: true38 w: 25639 ffn_w: 76840 n_slots: 1641 n_editor_blocks: 242 n_editor_heads: 443 r_max: 844 eta_max: 0.2545 clip_c: 1.046 clip_eps: 1.0e-0647 readout_zero_init: false48 readout_init_std: 0.249 readout_scale: 8.85156631469726650 controller_hidden: 051 ctrl_input: h52 ctrl_use_entropy: true53 lambda_cost: 0.054 kappa_risk: 0.055 dense_writes: false56