han1823123123 commited on
Commit
44397b3
·
verified ·
1 Parent(s): e851889

Upload folder using huggingface_hub

Browse files
01526d59abdd878c/config.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "act_cache_key": "98baddf64a997b15",
3
+ "agent": "unknown",
4
+ "arch": "stacked_sae",
5
+ "arch_version": "1.0.0",
6
+ "datasource": "toy_markov_n20_d40_noisy",
7
+ "saved_ts": "2026-05-06T13:39:45Z",
8
+ "seed": 42,
9
+ "train_key": "01526d59abdd878c",
10
+ "training_cfg": {
11
+ "arch_hparams_override": {
12
+ "T": 2,
13
+ "d_sae": 40,
14
+ "k_pos": 4
15
+ },
16
+ "batch_size": 1024,
17
+ "bricken_enabled": false,
18
+ "bricken_max_resample_fraction": 0.5,
19
+ "bricken_min_fires": 1,
20
+ "bricken_n_check": 2048,
21
+ "bricken_resample_every": 500,
22
+ "dead_threshold_tokens": 10000000,
23
+ "ema_auxk_alpha": 0.03125,
24
+ "learning_rate": 0.0003,
25
+ "n_steps": 30000,
26
+ "optimizer": "adam",
27
+ "plateau_early_stop": false,
28
+ "plateau_min_delta": 0.0001,
29
+ "plateau_patience": 5000,
30
+ "precision": "bf16",
31
+ "train_window_size": null,
32
+ "warmup_steps": 1000
33
+ }
34
+ }
01526d59abdd878c/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a0e468cb3cbda70e1fc5298e574ccdbbced899e171d709aa5c38a8804ecce960
3
+ size 26840