Abdullah
commited on
Uploading saes for ['transformer.h.0.attn', 'transformer.h.0.mlp', 'transformer.h.1.attn', 'transformer.h.1.mlp'] and sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32
Browse files- saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/model_config.json +1 -1
- saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/optimizer.pt +1 -1
- saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/state.pt +1 -1
- saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.0.attn/metrics.json +1 -1
- saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.0.attn/sae.safetensors +1 -1
- saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.0.mlp/metrics.json +1 -1
- saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.0.mlp/sae.safetensors +1 -1
- saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.1.attn/metrics.json +1 -1
- saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.1.attn/sae.safetensors +1 -1
- saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.1.mlp/metrics.json +1 -1
- saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.1.mlp/sae.safetensors +1 -1
saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/model_config.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"model_name": "withmartian/sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1", "dataset_name": "withmartian/
|
|
|
|
| 1 |
+
{"model_name": "withmartian/sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1", "dataset_name": "withmartian/cs5_dataset_synonyms", "model_abbrev": "saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True"}
|
saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1007171903
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ac45e9c96ef07c06f7caf5e08e5f381186af424482410a6c5ff310264b2821e6
|
| 3 |
size 1007171903
|
saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/state.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 493769
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1a21791269f0132fad465c7879f5c7465f6bca1ab53b1f601f3bcaff84edfa0b
|
| 3 |
size 493769
|
saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.0.attn/metrics.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"dead_pct": 0.
|
|
|
|
| 1 |
+
{"dead_pct": 0.4452474117279053, "fvu": 0.030083920806646347}
|
saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.0.attn/sae.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 125894984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:17150a5b9107576b97ea47372bd40fd143eb8c2c0f6fcc72675a228d2f168252
|
| 3 |
size 125894984
|
saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.0.mlp/metrics.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"dead_pct": 0.
|
|
|
|
| 1 |
+
{"dead_pct": 0.07851562649011612, "fvu": 0.05118158459663391}
|
saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.0.mlp/sae.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 125894984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dbbc7c87cba60665bed538de3452297c9d0a7458e140edd82ca19de63f861f28
|
| 3 |
size 125894984
|
saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.1.attn/metrics.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"dead_pct": 0.
|
|
|
|
| 1 |
+
{"dead_pct": 0.5203125476837158, "fvu": 0.052874982357025146}
|
saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.1.attn/sae.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 125894984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:95ced1033fd1128018871f32569f83f15f8b69f4c92f2682a2d2b0395a0b86dd
|
| 3 |
size 125894984
|
saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.1.mlp/metrics.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"dead_pct": 0.
|
|
|
|
| 1 |
+
{"dead_pct": 0.14557293057441711, "fvu": 0.061192333698272705}
|
saes_sft_sql_interp_TinyStories-2Layers-33M_cs4_dataset_synonyms_experiment_1.1_syn=True/k=32/transformer.h.1.mlp/sae.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 125894984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6af71d10ebbe039a91f4c8899d7de0bf27742c76cdcee7b6a7b71219b7ba46ab
|
| 3 |
size 125894984
|