{ "arm": "context", "data": "data/world-a", "out": "runs/context-a", "edits": 0, "mix": [ [ "qa", 1.0 ] ], "init": null, "init_checkpoint": "final", "size": "small", "dtype": "float32", "steps": 10000, "batch_size": 64, "seq_len": 256, "lr": 0.001, "min_lr_frac": 0.1, "warmup": 500, "weight_decay": 0.1, "grad_clip": 1.0, "log_every": 100, "eval_every": 2000, "eval_sets": "test_id,test_ood,test_1hop_ood", "eval_n": 200, "save_every": 2000, "seed": 0, "model": { "vocab_size": 576, "n_layer": 8, "n_head": 8, "d_model": 512, "mlp_ratio": 4, "rope_base": 10000.0 }, "params": 25469440, "vocab": 544 }