{ "measurement": { "model.layers.0": { "accuracy": 0.6285324096679688, "total_bits": 289320576, "o_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.1": { "accuracy": 0.695317268371582, "total_bits": 289320576, "o_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.2": { "accuracy": 0.7464818954467773, "total_bits": 283125504, "o_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.3": { "accuracy": 0.8897502422332764, "total_bits": 289320576, "o_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.4": { "accuracy": 0.7123856544494629, "total_bits": 301703808, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.5": { "accuracy": 0.7461767196655273, "total_bits": 233583360, "o_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.6": { "accuracy": 0.7727584838867188, "total_bits": 233583360, "o_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.7": { "accuracy": 0.798675537109375, "total_bits": 252161664, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.8": { "accuracy": 0.8624000549316406, "total_bits": 233583360, "o_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.9": { "accuracy": 0.8949522972106934, "total_bits": 252161664, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.10": { "accuracy": 0.8948664665222168, "total_bits": 252161664, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.11": { "accuracy": 0.8961267471313477, "total_bits": 252161664, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.12": { "accuracy": 0.8918008804321289, "total_bits": 252161664, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.13": { "accuracy": 0.9145417213439941, "total_bits": 283125504, "o_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.14": { "accuracy": 0.932328462600708, "total_bits": 233583360, "o_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.15": { "accuracy": 0.9327261447906494, "total_bits": 233583360, "o_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.16": { "accuracy": 0.9302468299865723, "total_bits": 233583360, "o_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.17": { "accuracy": 0.926159143447876, "total_bits": 233583360, "o_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.18": { "accuracy": 0.9283664226531982, "total_bits": 252161664, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.19": { "accuracy": 0.9305317401885986, "total_bits": 252161664, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.20": { "accuracy": 0.9183728694915771, "total_bits": 252161664, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.21": { "accuracy": 0.9167120456695557, "total_bits": 252161664, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.22": { "accuracy": 0.9127471446990967, "total_bits": 252161664, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.23": { "accuracy": 0.9099318981170654, "total_bits": 252161664, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.24": { "accuracy": 0.9029037952423096, "total_bits": 252161664, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.25": { "accuracy": 0.9138438701629639, "total_bits": 283125504, "o_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.26": { "accuracy": 0.9242894649505615, "total_bits": 301703808, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.27": { "accuracy": 0.921504020690918, "total_bits": 301703808, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.28": { "accuracy": 0.9472332000732422, "total_bits": 400793088, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "2": 64 }, "bits": [ 2 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.29": { "accuracy": 0.9694942235946655, "total_bits": 456530304, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.30": { "accuracy": 0.9758806824684143, "total_bits": 456530304, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 } }, "model.layers.31": { "accuracy": 0.9664642810821533, "total_bits": 456530304, "o_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "qkv_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 }, "gate_up_proj": { "group_size": { "4": 128 }, "bits": [ 4 ], "bits_prop": [ 1 ], "scale_bits": 4 } } } }