| layer,module,loss,samples,damp,time | |
| 0,linear_attn.in_proj_qkv,0.0000090902,0.05000,0.617 | |
| 0,linear_attn.in_proj_z,0.0000045542,0.05000,0.438 | |
| 0,linear_attn.out_proj,0.0000000044,0.05000,0.428 | |
| 0,mlp.up_proj,0.0000001226,0.05000,0.900 | |
| 0,mlp.gate_proj,0.0000001421,0.05000,0.892 | |
| 0,mlp.down_proj,0.0000000008,0.05000,1.595 | |
| 1,linear_attn.in_proj_qkv,0.0000011049,0.05000,0.507 | |
| 1,linear_attn.in_proj_z,0.0000005488,0.05000,0.443 | |
| 1,linear_attn.out_proj,0.0000000016,0.05000,0.443 | |
| 1,mlp.gate_proj,0.0000003634,0.05000,0.904 | |
| 1,mlp.up_proj,0.0000003239,0.05000,0.908 | |
| 1,mlp.down_proj,0.0000000023,0.05000,1.608 | |
| 2,linear_attn.in_proj_qkv,0.0000019753,0.05000,0.492 | |
| 2,linear_attn.in_proj_z,0.0000009680,0.05000,0.428 | |
| 2,linear_attn.out_proj,0.0000000029,0.05000,0.437 | |
| 2,mlp.gate_proj,0.0000006150,0.05000,0.874 | |
| 2,mlp.up_proj,0.0000005168,0.05000,0.876 | |
| 2,mlp.down_proj,0.0000000065,0.05000,1.617 | |
| 3,self_attn.k_proj,0.0000008796,0.05000,1.330 | |
| 3,self_attn.v_proj,0.0000008376,0.05000,1.333 | |
| 3,self_attn.q_proj,0.0000084286,0.05000,1.336 | |
| 3,self_attn.o_proj,0.0000000050,0.05000,0.445 | |
| 3,mlp.gate_proj,0.0000008091,0.05000,0.860 | |
| 3,mlp.up_proj,0.0000007321,0.05000,0.863 | |
| 3,mlp.down_proj,0.0000000084,0.05000,1.601 | |
| 4,linear_attn.in_proj_qkv,0.0000034649,0.05000,0.520 | |
| 4,linear_attn.in_proj_z,0.0000017468,0.05000,0.433 | |
| 4,linear_attn.out_proj,0.0000000080,0.05000,0.441 | |
| 4,mlp.up_proj,0.0000011547,0.05000,0.847 | |
| 4,mlp.gate_proj,0.0000012759,0.05000,0.852 | |
| 4,mlp.down_proj,0.0000000154,0.05000,1.577 | |
| 5,linear_attn.in_proj_qkv,0.0000047500,0.05000,0.516 | |
| 5,linear_attn.in_proj_z,0.0000022957,0.05000,0.443 | |
| 5,linear_attn.out_proj,0.0000000164,0.05000,0.444 | |
| 5,mlp.gate_proj,0.0000018376,0.05000,0.874 | |
| 5,mlp.up_proj,0.0000015494,0.05000,0.879 | |
| 5,mlp.down_proj,0.0000000277,0.05000,1.595 | |
| 6,linear_attn.in_proj_qkv,0.0000052777,0.05000,0.515 | |
| 6,linear_attn.in_proj_z,0.0000021524,0.05000,0.432 | |
| 6,linear_attn.out_proj,0.0000000324,0.05000,0.444 | |
| 6,mlp.gate_proj,0.0000025880,0.05000,0.863 | |
| 6,mlp.up_proj,0.0000021090,0.05000,0.867 | |
| 6,mlp.down_proj,0.0000000539,0.05000,1.594 | |
| 7,self_attn.v_proj,0.0000006785,0.05000,1.403 | |
| 7,self_attn.k_proj,0.0000007733,0.05000,1.403 | |
| 7,self_attn.q_proj,0.0000058339,0.05000,1.405 | |
| 7,self_attn.o_proj,0.0000000288,0.05000,0.437 | |
| 7,mlp.gate_proj,0.0000025658,0.05000,0.862 | |
| 7,mlp.up_proj,0.0000020487,0.05000,0.865 | |
| 7,mlp.down_proj,0.0000000413,0.05000,1.591 | |
| 8,linear_attn.in_proj_qkv,0.0000057309,0.05000,0.492 | |
| 8,linear_attn.in_proj_z,0.0000024447,0.05000,0.457 | |
| 8,linear_attn.out_proj,0.0000000340,0.05000,0.428 | |
| 8,mlp.gate_proj,0.0000026064,0.05000,0.845 | |
| 8,mlp.up_proj,0.0000021556,0.05000,0.848 | |
| 8,mlp.down_proj,0.0000000424,0.05000,1.598 | |
| 9,linear_attn.in_proj_qkv,0.0000059482,0.05000,0.518 | |
| 9,linear_attn.in_proj_z,0.0000024183,0.05000,0.438 | |
| 9,linear_attn.out_proj,0.0000000334,0.05000,0.440 | |
| 9,mlp.up_proj,0.0000022724,0.05000,0.928 | |
| 9,mlp.gate_proj,0.0000025724,0.05000,0.932 | |
| 9,mlp.down_proj,0.0000000452,0.05000,1.599 | |
| 10,linear_attn.in_proj_qkv,0.0000057265,0.05000,0.517 | |
| 10,linear_attn.in_proj_z,0.0000022270,0.05000,0.443 | |
| 10,linear_attn.out_proj,0.0000000355,0.05000,0.433 | |
| 10,mlp.up_proj,0.0000023992,0.05000,0.843 | |
| 10,mlp.gate_proj,0.0000024826,0.05000,0.846 | |
| 10,mlp.down_proj,0.0000000511,0.05000,1.602 | |
| 11,self_attn.v_proj,0.0000006117,0.05000,1.439 | |
| 11,self_attn.k_proj,0.0000006401,0.05000,1.442 | |
| 11,self_attn.q_proj,0.0000044447,0.05000,1.445 | |
| 11,self_attn.o_proj,0.0000000447,0.05000,0.442 | |
| 11,mlp.gate_proj,0.0000023887,0.05000,0.921 | |
| 11,mlp.up_proj,0.0000024465,0.05000,0.925 | |
| 11,mlp.down_proj,0.0000000502,0.05000,1.588 | |
| 12,linear_attn.in_proj_qkv,0.0000057455,0.05000,0.500 | |
| 12,linear_attn.in_proj_z,0.0000020781,0.05000,0.454 | |
| 12,linear_attn.out_proj,0.0000000449,0.05000,0.439 | |
| 12,mlp.up_proj,0.0000025431,0.05000,0.867 | |
| 12,mlp.gate_proj,0.0000023881,0.05000,0.870 | |
| 12,mlp.down_proj,0.0000000545,0.05000,1.589 | |
| 13,linear_attn.in_proj_qkv,0.0000063455,0.05000,0.545 | |
| 13,linear_attn.in_proj_z,0.0000023766,0.05000,0.445 | |
| 13,linear_attn.out_proj,0.0000000424,0.05000,0.448 | |
| 13,mlp.gate_proj,0.0000024403,0.05000,0.973 | |
| 13,mlp.up_proj,0.0000026470,0.05000,0.976 | |
| 13,mlp.down_proj,0.0000000573,0.05000,1.596 | |
| 14,linear_attn.in_proj_qkv,0.0000062900,0.05000,0.517 | |
| 14,linear_attn.in_proj_z,0.0000023549,0.05000,0.446 | |
| 14,linear_attn.out_proj,0.0000000682,0.05000,0.436 | |
| 14,mlp.up_proj,0.0000027956,0.05000,0.887 | |
| 14,mlp.gate_proj,0.0000024443,0.05000,0.889 | |
| 14,mlp.down_proj,0.0000000637,0.05000,1.598 | |
| 15,self_attn.k_proj,0.0000006541,0.05000,1.363 | |
| 15,self_attn.v_proj,0.0000008126,0.05000,1.366 | |
| 15,self_attn.q_proj,0.0000045815,0.05000,1.369 | |
| 15,self_attn.o_proj,0.0000000466,0.05000,0.448 | |
| 15,mlp.up_proj,0.0000030350,0.05000,0.861 | |
| 15,mlp.gate_proj,0.0000026180,0.05000,0.864 | |
| 15,mlp.down_proj,0.0000000735,0.05000,1.614 | |
| 16,linear_attn.in_proj_qkv,0.0000065196,0.05000,0.511 | |
| 16,linear_attn.in_proj_z,0.0000024188,0.05000,0.437 | |
| 16,linear_attn.out_proj,0.0000000654,0.05000,0.437 | |
| 16,mlp.gate_proj,0.0000028569,0.05000,0.846 | |
| 16,mlp.up_proj,0.0000032812,0.05000,0.849 | |
| 16,mlp.down_proj,0.0000000955,0.05000,1.592 | |
| 17,linear_attn.in_proj_qkv,0.0000087892,0.05000,0.520 | |
| 17,linear_attn.in_proj_z,0.0000026956,0.05000,0.454 | |
| 17,linear_attn.out_proj,0.0000000865,0.05000,0.442 | |
| 17,mlp.up_proj,0.0000036079,0.05000,0.921 | |
| 17,mlp.gate_proj,0.0000031055,0.05000,0.924 | |
| 17,mlp.down_proj,0.0000001271,0.05000,1.603 | |
| 18,linear_attn.in_proj_qkv,0.0000083640,0.05000,0.512 | |
| 18,linear_attn.in_proj_z,0.0000029537,0.05000,0.443 | |
| 18,linear_attn.out_proj,0.0000001388,0.05000,0.438 | |
| 18,mlp.gate_proj,0.0000040403,0.05000,0.863 | |
| 18,mlp.up_proj,0.0000044091,0.05000,0.865 | |
| 18,mlp.down_proj,0.0000002889,0.05000,1.606 | |
| 19,self_attn.k_proj,0.0000009482,0.05000,1.410 | |
| 19,self_attn.q_proj,0.0000062437,0.05000,1.414 | |
| 19,self_attn.v_proj,0.0000020255,0.05000,1.420 | |
| 19,self_attn.o_proj,0.0000001750,0.05000,0.449 | |
| 19,mlp.up_proj,0.0000050957,0.05000,0.919 | |
| 19,mlp.gate_proj,0.0000044884,0.05000,0.925 | |
| 19,mlp.down_proj,0.0000003516,0.05000,1.601 | |
| 20,linear_attn.in_proj_qkv,0.0000128556,0.05000,0.508 | |
| 20,linear_attn.in_proj_z,0.0000044074,0.05000,0.435 | |
| 20,linear_attn.out_proj,0.0000001916,0.05000,0.448 | |
| 20,mlp.gate_proj,0.0000056739,0.05000,0.850 | |
| 20,mlp.up_proj,0.0000054167,0.05000,0.852 | |
| 20,mlp.down_proj,0.0000004084,0.05000,1.583 | |
| 21,linear_attn.in_proj_qkv,0.0000125873,0.05000,0.521 | |
| 21,linear_attn.in_proj_z,0.0000045462,0.05000,0.452 | |
| 21,linear_attn.out_proj,0.0000002626,0.05000,0.440 | |
| 21,mlp.up_proj,0.0000057516,0.05000,0.868 | |
| 21,mlp.gate_proj,0.0000064190,0.05000,0.868 | |
| 21,mlp.down_proj,0.0000005064,0.05000,1.589 | |
| 22,linear_attn.in_proj_qkv,0.0000123458,0.05000,0.515 | |
| 22,linear_attn.in_proj_z,0.0000051228,0.05000,0.448 | |
| 22,linear_attn.out_proj,0.0000002861,0.05000,0.447 | |
| 22,mlp.up_proj,0.0000073939,0.05000,0.979 | |
| 22,mlp.gate_proj,0.0000091414,0.05000,0.983 | |
| 22,mlp.down_proj,0.0000009588,0.05000,1.628 | |
| 23,self_attn.q_proj,0.0000083992,0.05000,1.383 | |
| 23,self_attn.v_proj,0.0000027723,0.05000,1.386 | |
| 23,self_attn.k_proj,0.0000013828,0.05000,1.387 | |
| 23,self_attn.o_proj,0.0000002685,0.05000,0.459 | |
| 23,mlp.gate_proj,0.0000081927,0.05000,0.883 | |
| 23,mlp.up_proj,0.0000072828,0.05000,0.889 | |
| 23,mlp.down_proj,0.0000008112,0.05000,1.607 | |
| 24,linear_attn.in_proj_qkv,0.0000150573,0.05000,0.511 | |
| 24,linear_attn.in_proj_z,0.0000055410,0.05000,0.443 | |
| 24,linear_attn.out_proj,0.0000005550,0.05000,0.452 | |
| 24,mlp.gate_proj,0.0000091673,0.05000,0.875 | |
| 24,mlp.up_proj,0.0000077243,0.05000,0.876 | |
| 24,mlp.down_proj,0.0000008447,0.05000,1.599 | |
| 25,linear_attn.in_proj_qkv,0.0000159640,0.05000,0.521 | |
| 25,linear_attn.in_proj_z,0.0000059781,0.05000,0.447 | |
| 25,linear_attn.out_proj,0.0000004635,0.05000,0.440 | |
| 25,mlp.gate_proj,0.0000107524,0.05000,0.874 | |
| 25,mlp.up_proj,0.0000087090,0.05000,0.877 | |
| 25,mlp.down_proj,0.0000008964,0.05000,1.595 | |
| 26,linear_attn.in_proj_qkv,0.0000145905,0.05000,0.532 | |
| 26,linear_attn.in_proj_z,0.0000062938,0.05000,0.447 | |
| 26,linear_attn.out_proj,0.0000005024,0.05000,0.445 | |
| 26,mlp.gate_proj,0.0000126227,0.05000,0.883 | |
| 26,mlp.up_proj,0.0000100352,0.05000,0.885 | |
| 26,mlp.down_proj,0.0000013361,0.05000,1.590 | |
| 27,self_attn.k_proj,0.0000018342,0.05000,1.449 | |
| 27,self_attn.q_proj,0.0000097237,0.05000,1.453 | |
| 27,self_attn.v_proj,0.0000066289,0.05000,1.459 | |
| 27,self_attn.o_proj,0.0000011969,0.05000,0.441 | |
| 27,mlp.gate_proj,0.0000124524,0.05000,0.865 | |
| 27,mlp.up_proj,0.0000101295,0.05000,0.867 | |
| 27,mlp.down_proj,0.0000013725,0.05000,1.599 | |
| 28,linear_attn.in_proj_qkv,0.0000180177,0.05000,0.503 | |
| 28,linear_attn.in_proj_z,0.0000067820,0.05000,0.441 | |
| 28,linear_attn.out_proj,0.0000011778,0.05000,0.460 | |
| 28,mlp.gate_proj,0.0000131360,0.05000,0.855 | |
| 28,mlp.up_proj,0.0000108148,0.05000,0.858 | |
| 28,mlp.down_proj,0.0000017392,0.05000,1.582 | |
| 29,linear_attn.in_proj_qkv,0.0000132031,0.05000,0.521 | |
| 29,linear_attn.in_proj_z,0.0000060368,0.05000,0.446 | |
| 29,linear_attn.out_proj,0.0000010770,0.05000,0.447 | |
| 29,mlp.gate_proj,0.0000146661,0.05000,0.867 | |
| 29,mlp.up_proj,0.0000120007,0.05000,0.869 | |
| 29,mlp.down_proj,0.0000022625,0.05000,1.594 | |
| 30,linear_attn.in_proj_qkv,0.0000145864,0.05000,0.529 | |
| 30,linear_attn.in_proj_z,0.0000062206,0.05000,0.441 | |
| 30,linear_attn.out_proj,0.0000024084,0.05000,0.441 | |
| 30,mlp.gate_proj,0.0000138783,0.05000,0.876 | |
| 30,mlp.up_proj,0.0000115754,0.05000,0.882 | |
| 30,mlp.down_proj,0.0000031752,0.05000,1.556 | |
| 31,self_attn.k_proj,0.0000019336,0.05000,1.384 | |
| 31,self_attn.v_proj,0.0000060194,0.05000,1.387 | |
| 31,self_attn.q_proj,0.0000101493,0.05000,1.394 | |
| 31,self_attn.o_proj,0.0000037437,0.05000,0.445 | |
| 31,mlp.up_proj,0.0000104450,0.05000,0.883 | |
| 31,mlp.gate_proj,0.0000128998,0.05000,0.886 | |
| 31,mlp.down_proj,0.0000075799,0.05000,1.644 | |