petra345's picture
Release step_900 with plateau-aware scorecard
1a630b7 verified
Raw
History Blame Contribute Delete
1.56 kB
checkpoint,overall,delta_from_previous,passes_release_gate,selected,math_reasoning,logical_reasoning,common_sense,reading_comprehension,question_answering,text_classification,sentiment_analysis,code_generation,creative_writing,dialogue_generation,summarization,translation,knowledge_retrieval,instruction_following,safety_evaluation
step_100,0.480,,false,false,0.345,0.319,0.530,0.475,0.475,0.517,0.617,0.350,0.328,0.438,0.517,0.640,0.529,0.550,0.628
step_200,0.535,0.055,false,false,0.383,0.375,0.583,0.529,0.510,0.603,0.675,0.421,0.388,0.493,0.600,0.700,0.570,0.610,0.650
step_300,0.576,0.041,false,false,0.415,0.445,0.621,0.569,0.533,0.667,0.710,0.475,0.436,0.530,0.650,0.733,0.596,0.650,0.668
step_400,0.608,0.032,false,false,0.443,0.525,0.650,0.600,0.550,0.714,0.733,0.517,0.475,0.557,0.683,0.755,0.615,0.679,0.683
step_500,0.635,0.027,false,false,0.467,0.605,0.672,0.625,0.564,0.750,0.750,0.550,0.507,0.579,0.707,0.769,0.631,0.700,0.696
step_600,0.656,0.021,false,false,0.487,0.675,0.690,0.645,0.575,0.776,0.762,0.577,0.534,0.596,0.725,0.780,0.643,0.717,0.707
step_700,0.674,0.018,false,false,0.506,0.731,0.705,0.663,0.584,0.795,0.772,0.600,0.557,0.611,0.739,0.788,0.653,0.730,0.717
step_800,0.689,0.015,false,false,0.522,0.773,0.717,0.677,0.593,0.809,0.780,0.619,0.577,0.624,0.750,0.795,0.662,0.741,0.725
step_900,0.700,0.011,true,true,0.537,0.801,0.727,0.689,0.600,0.820,0.786,0.636,0.595,0.634,0.759,0.800,0.670,0.750,0.732
step_1000,0.710,0.010,true,false,0.550,0.819,0.736,0.700,0.607,0.828,0.792,0.650,0.610,0.644,0.767,0.804,0.676,0.758,0.739