Upload 2 files
Browse files- Config.json +36 -0
- SAE.pt +3 -0
Config.json
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"seed": 49,
|
| 3 |
+
"epochs": 5000,
|
| 4 |
+
"batch_size": 16384,
|
| 5 |
+
"lr": 0.001,
|
| 6 |
+
"l1_coeff": 0,
|
| 7 |
+
"beta1": 0.9,
|
| 8 |
+
"beta2": 0.99,
|
| 9 |
+
"num_tokens": 1000000000,
|
| 10 |
+
"max_grad_norm": 100000,
|
| 11 |
+
"seq_len": 128,
|
| 12 |
+
"dtype": "torch.float32",
|
| 13 |
+
"model_name": "gpt2-small",
|
| 14 |
+
"site": "resid_pre",
|
| 15 |
+
"layer": 8,
|
| 16 |
+
"act_size": 768,
|
| 17 |
+
"dict_size": 2304,
|
| 18 |
+
"device": "cuda:0",
|
| 19 |
+
"model_batch_size": 512,
|
| 20 |
+
"num_batches_in_buffer": 10,
|
| 21 |
+
"dataset_path": "Skylion007/openwebtext",
|
| 22 |
+
"wandb_project": "gpt2-feature-splitting-saes",
|
| 23 |
+
"input_unit_norm": false,
|
| 24 |
+
"perf_log_freq": 1000,
|
| 25 |
+
"sae_type": "topk",
|
| 26 |
+
"architecture": "standard",
|
| 27 |
+
"checkpoint_freq": 1000,
|
| 28 |
+
"n_batches_to_dead": 100,
|
| 29 |
+
"top_k": 4,
|
| 30 |
+
"top_k_aux": 512,
|
| 31 |
+
"aux_penalty": 0.0625,
|
| 32 |
+
"hook_point": "blocks.8.hook_resid_pre",
|
| 33 |
+
"threshold": null,
|
| 34 |
+
"name": "gpt2-small_blocks.8.hook_resid_pre_2304_topk_4_0.001",
|
| 35 |
+
"cosine_penalty": 0.0
|
| 36 |
+
}
|
SAE.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:569a630414d5176e7bdc62c8ee8fc521c53fa32ae34007274528ea9c1b109bf5
|
| 3 |
+
size 14169840
|