Upload folder using huggingface_hub
Browse files
s0/b0_residpost_sep_jumprelu_d2048_l016/trainer_0/ae.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d8a754dd58c20a8e4fd7597d79b0197ce6dab2958320f963c6e7f81cc57d05f3
|
| 3 |
+
size 4214394
|
s0/b0_residpost_sep_jumprelu_d2048_l016/trainer_0/config.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"trainer": {
|
| 3 |
+
"trainer_class": "JumpReluTrainer",
|
| 4 |
+
"dict_class": "JumpReluAutoEncoder",
|
| 5 |
+
"lr": 0.0001,
|
| 6 |
+
"steps": 5000000,
|
| 7 |
+
"seed": 0,
|
| 8 |
+
"activation_dim": 256,
|
| 9 |
+
"dict_size": 2048,
|
| 10 |
+
"device": "cuda",
|
| 11 |
+
"layer": 0,
|
| 12 |
+
"lm_name": "toy_binding",
|
| 13 |
+
"wandb_name": "b0_residpost_sep_jumprelu_d2048_l016",
|
| 14 |
+
"submodule_name": null,
|
| 15 |
+
"bandwidth": 0.001,
|
| 16 |
+
"sparsity_penalty": 1.0,
|
| 17 |
+
"sparsity_warmup_steps": 2000,
|
| 18 |
+
"target_l0": 16.0
|
| 19 |
+
}
|
| 20 |
+
}
|