Upload training_logs/txcdr_t28__seed1.json with huggingface_hub
Browse files
training_logs/txcdr_t28__seed1.json
ADDED
|
@@ -0,0 +1,105 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"loss": [
|
| 3 |
+
127048.21875,
|
| 4 |
+
9010.2509765625,
|
| 5 |
+
8322.990234375,
|
| 6 |
+
8019.96240234375,
|
| 7 |
+
7647.80712890625,
|
| 8 |
+
7438.34814453125,
|
| 9 |
+
7287.4091796875,
|
| 10 |
+
7144.302734375,
|
| 11 |
+
7087.75927734375,
|
| 12 |
+
6982.732421875,
|
| 13 |
+
6938.48486328125,
|
| 14 |
+
6905.73486328125,
|
| 15 |
+
6874.05615234375,
|
| 16 |
+
6854.83154296875,
|
| 17 |
+
6804.0595703125,
|
| 18 |
+
6737.779296875,
|
| 19 |
+
6765.11083984375,
|
| 20 |
+
6732.81982421875,
|
| 21 |
+
6722.30810546875,
|
| 22 |
+
6704.28662109375,
|
| 23 |
+
6705.99267578125
|
| 24 |
+
],
|
| 25 |
+
"l0": [
|
| 26 |
+
500.0,
|
| 27 |
+
496.436767578125,
|
| 28 |
+
495.828369140625,
|
| 29 |
+
496.11669921875,
|
| 30 |
+
495.3564453125,
|
| 31 |
+
495.84716796875,
|
| 32 |
+
495.854248046875,
|
| 33 |
+
496.427490234375,
|
| 34 |
+
496.763427734375,
|
| 35 |
+
494.985107421875,
|
| 36 |
+
495.532470703125,
|
| 37 |
+
496.35205078125,
|
| 38 |
+
496.461181640625,
|
| 39 |
+
496.23974609375,
|
| 40 |
+
495.359375,
|
| 41 |
+
495.136474609375,
|
| 42 |
+
496.354248046875,
|
| 43 |
+
495.443115234375,
|
| 44 |
+
496.313720703125,
|
| 45 |
+
495.83984375,
|
| 46 |
+
494.866455078125
|
| 47 |
+
],
|
| 48 |
+
"steps_logged": [
|
| 49 |
+
0,
|
| 50 |
+
200,
|
| 51 |
+
400,
|
| 52 |
+
600,
|
| 53 |
+
800,
|
| 54 |
+
1000,
|
| 55 |
+
1200,
|
| 56 |
+
1400,
|
| 57 |
+
1600,
|
| 58 |
+
1800,
|
| 59 |
+
2000,
|
| 60 |
+
2200,
|
| 61 |
+
2400,
|
| 62 |
+
2600,
|
| 63 |
+
2800,
|
| 64 |
+
3000,
|
| 65 |
+
3200,
|
| 66 |
+
3400,
|
| 67 |
+
3600,
|
| 68 |
+
3800,
|
| 69 |
+
4000
|
| 70 |
+
],
|
| 71 |
+
"final_step": 4000,
|
| 72 |
+
"converged": true,
|
| 73 |
+
"plateau_last": 0.015974250598195773,
|
| 74 |
+
"elapsed_s": 4118.382341384888,
|
| 75 |
+
"row": 28,
|
| 76 |
+
"arch_id": "txcdr_t28",
|
| 77 |
+
"arch": "txcdr_t28",
|
| 78 |
+
"group": 3,
|
| 79 |
+
"src_class": "TemporalCrosscoder",
|
| 80 |
+
"src_module": "src.architectures.crosscoder",
|
| 81 |
+
"T": 28,
|
| 82 |
+
"T_max": null,
|
| 83 |
+
"t_sample": null,
|
| 84 |
+
"n_layers": null,
|
| 85 |
+
"k_win": 500,
|
| 86 |
+
"k_pos": 18,
|
| 87 |
+
"shifts": null,
|
| 88 |
+
"alpha": null,
|
| 89 |
+
"gamma": null,
|
| 90 |
+
"n_scales": null,
|
| 91 |
+
"seed": 1,
|
| 92 |
+
"d_in": 2304,
|
| 93 |
+
"d_sae": 18432,
|
| 94 |
+
"subject_model": "google/gemma-2-2b",
|
| 95 |
+
"anchor_layer": 12,
|
| 96 |
+
"mlc_layers": [
|
| 97 |
+
10,
|
| 98 |
+
11,
|
| 99 |
+
12,
|
| 100 |
+
13,
|
| 101 |
+
14
|
| 102 |
+
],
|
| 103 |
+
"phase": "phase7_unification",
|
| 104 |
+
"run_id": "txcdr_t28__seed1"
|
| 105 |
+
}
|