Upload folder using huggingface_hub
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- config.json +1 -0
- model.layers.0.mlp/cfg.json +1 -0
- model.layers.0.mlp/sae.safetensors +3 -0
- model.layers.1.mlp/cfg.json +1 -0
- model.layers.1.mlp/sae.safetensors +3 -0
- model.layers.10.mlp/cfg.json +1 -0
- model.layers.10.mlp/sae.safetensors +3 -0
- model.layers.11.mlp/cfg.json +1 -0
- model.layers.11.mlp/sae.safetensors +3 -0
- model.layers.12.mlp/cfg.json +1 -0
- model.layers.12.mlp/sae.safetensors +3 -0
- model.layers.13.mlp/cfg.json +1 -0
- model.layers.13.mlp/sae.safetensors +3 -0
- model.layers.14.mlp/cfg.json +1 -0
- model.layers.14.mlp/sae.safetensors +3 -0
- model.layers.15.mlp/cfg.json +1 -0
- model.layers.15.mlp/sae.safetensors +3 -0
- model.layers.16.mlp/cfg.json +1 -0
- model.layers.16.mlp/sae.safetensors +3 -0
- model.layers.17.mlp/cfg.json +1 -0
- model.layers.17.mlp/sae.safetensors +3 -0
- model.layers.18.mlp/cfg.json +1 -0
- model.layers.18.mlp/sae.safetensors +3 -0
- model.layers.19.mlp/cfg.json +1 -0
- model.layers.19.mlp/sae.safetensors +3 -0
- model.layers.2.mlp/cfg.json +1 -0
- model.layers.2.mlp/sae.safetensors +3 -0
- model.layers.20.mlp/cfg.json +1 -0
- model.layers.20.mlp/sae.safetensors +3 -0
- model.layers.21.mlp/cfg.json +1 -0
- model.layers.21.mlp/sae.safetensors +3 -0
- model.layers.22.mlp/cfg.json +1 -0
- model.layers.22.mlp/sae.safetensors +3 -0
- model.layers.23.mlp/cfg.json +1 -0
- model.layers.23.mlp/sae.safetensors +3 -0
- model.layers.24.mlp/cfg.json +1 -0
- model.layers.24.mlp/sae.safetensors +3 -0
- model.layers.26.mlp/cfg.json +1 -0
- model.layers.26.mlp/sae.safetensors +3 -0
- model.layers.27.mlp/cfg.json +1 -0
- model.layers.27.mlp/sae.safetensors +3 -0
- model.layers.28.mlp/cfg.json +1 -0
- model.layers.28.mlp/sae.safetensors +3 -0
- model.layers.29.mlp/cfg.json +1 -0
- model.layers.29.mlp/sae.safetensors +3 -0
- model.layers.3.mlp/cfg.json +1 -0
- model.layers.3.mlp/sae.safetensors +3 -0
- model.layers.4.mlp/cfg.json +1 -0
- model.layers.4.mlp/sae.safetensors +3 -0
- model.layers.5.mlp/cfg.json +1 -0
config.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"sae": {"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true}, "batch_size": 8, "grad_acc_steps": 4, "micro_acc_steps": 1, "optimizer": "signum", "lr": null, "lr_warmup_steps": 1000, "k_decay_steps": 0, "auxk_alpha": 0.0, "dead_feature_threshold": 10000000, "hookpoints": ["model.layers.8.mlp"], "init_seeds": [0], "layers": [], "layer_stride": 1, "distribute_modules": false, "save_every": 1000, "log_to_wandb": true, "run_name": "single_128x", "wandb_log_frequency": 1, "model": "HuggingFaceTB/SmolLM2-135M", "dataset": "EleutherAI/fineweb-edu-dedup-10b", "split": "train", "ctx_len": 2048, "hf_token": null, "revision": null, "load_in_8bit": false, "max_examples": 500000, "resume": false, "text_column": "text", "finetune": null, "shuffle_seed": 42, "data_preprocessing_num_proc": 48}
|
model.layers.0.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.0.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:56ce70caa3f7167f8aefc3a2ba6cc0b782795a29f15f5ac21f0caaf5b7fa2677
|
3 |
+
size 341363360
|
model.layers.1.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.1.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2a0bd89cde09b2b4f9b0e48726cdc7e4cb20337840d9281354946f3c11a4b3ad
|
3 |
+
size 341363360
|
model.layers.10.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.10.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5c0b4b58019665f4358c54f451b11a41c34cd48c2fc78b4e0eff730455fc57d7
|
3 |
+
size 341363360
|
model.layers.11.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.11.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:616287eec940898fc3cb72ef2bb2d09b6a5cebd6256217b197bcb5058cae92f9
|
3 |
+
size 341363360
|
model.layers.12.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.12.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3d6c03f54af342471bfdb8a46b7707c2839a320d30a93d1ba6506e16ee9a4e43
|
3 |
+
size 341363360
|
model.layers.13.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.13.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c02de3b5882e3053cf5d75b703a63ed8b0ddcc8843bb3305adf8ab2252a17672
|
3 |
+
size 341363360
|
model.layers.14.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.14.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:55a455c3c2350aec1824ad9e8041305c7a6011196730f3ef0e104647eb59d9a3
|
3 |
+
size 341363360
|
model.layers.15.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.15.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:93b1262ece6ba839c2825649f84ef7a1264038f5c561472e1d8c7665a3cf42a1
|
3 |
+
size 341363360
|
model.layers.16.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.16.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1b4207e91baeacb6094a314b995b29886fcbf82fbf90dfbe5effc90fe5bf3cfb
|
3 |
+
size 341363360
|
model.layers.17.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.17.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:837099fd7dc4f9b36da8b403fe410fb85bc03f4343160057770fe7a220101eab
|
3 |
+
size 341363360
|
model.layers.18.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.18.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fad15c681e948baafe0bb48ddc66e14910f17967825dfcff070fb5a6132e6a91
|
3 |
+
size 341363360
|
model.layers.19.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.19.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d720fd8ebfaf1d66abf36d82b7249e89dbe3c9bfc2c01228885576babd9ee7f0
|
3 |
+
size 341363360
|
model.layers.2.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.2.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7f981a5e57d2e40cf9601057809d72909fa8fed78fcae3775f4f8f72c423fc11
|
3 |
+
size 341363360
|
model.layers.20.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.20.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7bbf6f1dc5377a17525fb999c4b10897c86ced205579c37121c8d6dc735b06ac
|
3 |
+
size 341363360
|
model.layers.21.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.21.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a89aa886e34339841f1d18aca73a4d07800d3628680f894ad5c322d924be86b0
|
3 |
+
size 341363360
|
model.layers.22.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.22.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d176fb7a5a2688cad50a9c22cfcebb8cd1e0972639dc79f9e77a2fa49aed2750
|
3 |
+
size 341363360
|
model.layers.23.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.23.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:69df70a0918bb431f9e4d3b2bbdc80d48af11c94eb8f367520142d94c551fd42
|
3 |
+
size 341363360
|
model.layers.24.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.24.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d4e39c2c700f3f31a99e4ac0a2751176f5f2472c1c48e28bdcf0279314605c92
|
3 |
+
size 341363360
|
model.layers.26.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true,"d_in": 576}
|
model.layers.26.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:430a72cf5fa68b97b887b93b289ee007353db785d04971fe64059edc02b703ab
|
3 |
+
size 341363360
|
model.layers.27.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.27.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8317d1b216779c8667832b55664fabfe08fab6c3444856ce096c980d1b3bddb1
|
3 |
+
size 341363360
|
model.layers.28.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.28.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cb59358c254fbebfba816bdb06a6fc508b83c189d4a3d3eb7569b7d3f9f19cea
|
3 |
+
size 341363360
|
model.layers.29.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.29.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bb725eee13d13ffcc495640d082ff12699e056d3af2dcad279597288bfd52670
|
3 |
+
size 341363360
|
model.layers.3.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.3.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:842613d636debd426d83fc6f8f0e2eaa4a788d28fdd6305a996e862fb6d67e0b
|
3 |
+
size 341363360
|
model.layers.4.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|
model.layers.4.mlp/sae.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d3d7788cac85144c347d4f49f5706b1be856756e92c58f32791777afbb2bd1a7
|
3 |
+
size 341363360
|
model.layers.5.mlp/cfg.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"activation": "topk", "expansion_factor": 128, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": true, "transcode": true, "d_in": 576}
|