{ "overall": { "sparsity": 0.3300018310546875, "params": 6979321856, "zero_params": 2303188992, "remaining_params": 4676132864, "delta_sparsity": 0.09568375807542068, "params_pruned_this_step": 667807744 }, "by_type": { "attn_out": { "layers": 32, "sparsity": 0.3300018310546875, "zero_rows": 0, "rows": 131072, "zero_cols": 43254, "cols": 131072 }, "ff_out": { "layers": 32, "sparsity": 0.3300018310546875, "zero_rows": 0, "rows": 131072, "zero_cols": 129762, "cols": 393216 }, "ff_proj": { "layers": 32, "sparsity": 0.3300018310546875, "zero_rows": 129762, "rows": 393216, "zero_cols": 0, "cols": 131072 }, "k_proj": { "layers": 32, "sparsity": 0.3300018310546875, "zero_rows": 43254, "rows": 131072, "zero_cols": 0, "cols": 131072 }, "q_proj": { "layers": 32, "sparsity": 0.3300018310546875, "zero_rows": 43254, "rows": 131072, "zero_cols": 0, "cols": 131072 }, "up_proj": { "layers": 32, "sparsity": 0.3300018310546875, "zero_rows": 129762, "rows": 393216, "zero_cols": 0, "cols": 131072 }, "v_proj": { "layers": 32, "sparsity": 0.3300018310546875, "zero_rows": 43254, "rows": 131072, "zero_cols": 0, "cols": 131072 } }, "by_block": { "0": 0.8824556790865384, "1": 0.8342848557692307, "2": 0.7604792668269231, "3": 0.6421837439903846, "4": 0.4834172175480769, "5": 0.33486703725961536, "6": 0.3531963641826923, "7": 0.2225623497596154, "8": 0.25522085336538464, "9": 0.29604867788461536, "10": 0.30434945913461536, "11": 0.2946965144230769, "12": 0.2802734375, "13": 0.20870267427884615, "14": 0.2584698016826923, "15": 0.26185021033653844, "16": 0.2237830528846154, "17": 0.22705078125, "18": 0.24859149639423078, "19": 0.19844876802884615, "20": 0.20344426081730768, "21": 0.18776292067307693, "22": 0.17687049278846154, "23": 0.1517615685096154, "24": 0.12180739182692307, "25": 0.2080078125, "26": 0.23313551682692307, "27": 0.32070688100961536, "28": 0.38609900841346156, "29": 0.3822678786057692, "30": 0.3446702223557692, "31": 0.27259239783653844 } }