script_1/flexitok--bpe_script_Roma_32000_overlap.json

{"1": {"ratio_to_total_tokens": 0.6736941376897569, "expected_training_ratio_in_superset": 0.061244921608159725, "num_tokens": 22145}, "2": {"ratio_to_total_tokens": 0.10608134830093395, "expected_training_ratio_in_superset": 0.01928751787289708, "num_tokens": 3487}, "3": {"ratio_to_total_tokens": 0.06510297830914788, "expected_training_ratio_in_superset": 0.017755357720676695, "num_tokens": 2140}, "4": {"ratio_to_total_tokens": 0.04055246265705333, "expected_training_ratio_in_superset": 0.014746350057110303, "num_tokens": 1333}, "5": {"ratio_to_total_tokens": 0.025158954701712755, "expected_training_ratio_in_superset": 0.011435888500778525, "num_tokens": 827}, "6": {"ratio_to_total_tokens": 0.01627574457728697, "expected_training_ratio_in_superset": 0.008877678860338347, "num_tokens": 535}, "7": {"ratio_to_total_tokens": 0.012229624897325911, "expected_training_ratio_in_superset": 0.007782488571025579, "num_tokens": 402}, "8": {"ratio_to_total_tokens": 0.0069666271181284415, "expected_training_ratio_in_superset": 0.005066637904093413, "num_tokens": 229}, "9": {"ratio_to_total_tokens": 0.0026771318183200997, "expected_training_ratio_in_superset": 0.0021903805786255364, "num_tokens": 88}, "10": {"ratio_to_total_tokens": 0.0029205074381673814, "expected_training_ratio_in_superset": 0.0026550067619703467, "num_tokens": 96}, "11": {"ratio_to_total_tokens": 0.048340482492166346, "expected_training_ratio_in_superset": 0.048340482492166346, "num_tokens": 1589}, "total_training_compared_to_full_model": 0.1993827109278419}