Upload folder using huggingface_hub

config.json

@@ -0,0 +1 @@
+{"sae": {"expansion_factor": 32, "normalize_decoder": true, "num_latents": 768, "k": 32, "multi_topk": false, "skip_connection": false}, "batch_size": 8, "grad_acc_steps": 1, "micro_acc_steps": 1, "lr": null, "lr_warmup_steps": 1000, "auxk_alpha": 0.0, "dead_feature_threshold": 10000000, "hookpoints": ["layers.0.mlp", "layers.1.mlp", "layers.2.mlp", "layers.3.mlp", "layers.4.mlp", "layers.5.mlp", "layers.6.mlp", "layers.7.mlp", "layers.8.mlp", "layers.9.mlp", "layers.10.mlp", "layers.11.mlp"], "layers": [], "layer_stride": 1, "transcode": false, "distribute_modules": false, "save_every": 1000, "log_to_wandb": true, "run_name": "k32-sae-mlp-768", "wandb_log_frequency": 1, "model": "EleutherAI/pythia-160m", "dataset": "/mnt/ssd-1/pile_preshuffled/standard/document.bin", "split": "train", "ctx_len": 2049, "hf_token": null, "revision": null, "load_in_8bit": false, "max_examples": 4000000, "resume": false, "finetune": null, "seed": 42, "data_preprocessing_num_proc": 48}

layers.0.mlp/cfg.json

@@ -0,0 +1 @@
+{"expansion_factor": 32, "normalize_decoder": true, "num_latents": 768, "k": 32, "multi_topk": false, "skip_connection": false, "d_in": 768}

layers.0.mlp/sae.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7614a43b3ae994004a221c82b9e60d61fb71848f9321ec55ac2ab74a5488488
+size 4725048

layers.1.mlp/cfg.json

@@ -0,0 +1 @@
+{"expansion_factor": 32, "normalize_decoder": true, "num_latents": 768, "k": 32, "multi_topk": false, "skip_connection": false, "d_in": 768}

layers.1.mlp/sae.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2515f319716d0dfec91bc2182e7b529df72783c4f2ead97dbd6d80fd3935fa4b
+size 4725048

layers.10.mlp/cfg.json

@@ -0,0 +1 @@
+{"expansion_factor": 32, "normalize_decoder": true, "num_latents": 768, "k": 32, "multi_topk": false, "skip_connection": false, "d_in": 768}

layers.10.mlp/sae.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f637f55e7ec69a6987d519c9bd33b7952006029eb150a42323ccfeb5bd37f1a5
+size 4725048

layers.11.mlp/cfg.json

@@ -0,0 +1 @@
+{"expansion_factor": 32, "normalize_decoder": true, "num_latents": 768, "k": 32, "multi_topk": false, "skip_connection": false, "d_in": 768}

layers.11.mlp/sae.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1d5b7cdbfd9461063b46fc6f3c857e0855789f298eb6d4d559a60aba36c1ce4
+size 4725048

layers.2.mlp/cfg.json

@@ -0,0 +1 @@
+{"expansion_factor": 32, "normalize_decoder": true, "num_latents": 768, "k": 32, "multi_topk": false, "skip_connection": false, "d_in": 768}

layers.2.mlp/sae.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3307bde486a508327781dbd374c2871c9c20fc76fbe8a00b7ad7c46285165170
+size 4725048

layers.3.mlp/cfg.json

@@ -0,0 +1 @@
+{"expansion_factor": 32, "normalize_decoder": true, "num_latents": 768, "k": 32, "multi_topk": false, "skip_connection": false, "d_in": 768}

layers.3.mlp/sae.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0911261b3649db5831e0ad207960bc5d26de87ef2856ccf1b9bc1636dd26d53f
+size 4725048

layers.4.mlp/cfg.json

@@ -0,0 +1 @@
+{"expansion_factor": 32, "normalize_decoder": true, "num_latents": 768, "k": 32, "multi_topk": false, "skip_connection": false, "d_in": 768}

layers.4.mlp/sae.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23f882c21ee35e0730a0aaa0d06b394f205f4aa382d570978a57697bab7bab64
+size 4725048

layers.5.mlp/cfg.json

@@ -0,0 +1 @@
+{"expansion_factor": 32, "normalize_decoder": true, "num_latents": 768, "k": 32, "multi_topk": false, "skip_connection": false, "d_in": 768}

layers.5.mlp/sae.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:30d841d56cd246d8ae673f6748069db2e65e26ad774b385ff6a4a1152ca1a248
+size 4725048

layers.6.mlp/cfg.json

@@ -0,0 +1 @@
+{"expansion_factor": 32, "normalize_decoder": true, "num_latents": 768, "k": 32, "multi_topk": false, "skip_connection": false, "d_in": 768}

layers.6.mlp/sae.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1cc88ed824c27a8d5255f3f9193516b802d7da6d3aef8725d4a055b9dcc9a6dd
+size 4725048

layers.7.mlp/cfg.json

@@ -0,0 +1 @@
+{"expansion_factor": 32, "normalize_decoder": true, "num_latents": 768, "k": 32, "multi_topk": false, "skip_connection": false, "d_in": 768}

layers.7.mlp/sae.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9109097c677fa6b978f17e16b622adfb78328ac8d1f847f58e56b84ef5f412ee
+size 4725048

layers.8.mlp/cfg.json

@@ -0,0 +1 @@
+{"expansion_factor": 32, "normalize_decoder": true, "num_latents": 768, "k": 32, "multi_topk": false, "skip_connection": false, "d_in": 768}

layers.8.mlp/sae.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10d817e80557517da5413966d76e5664efbdea933ed9f2f839fb02a1d2397a58
+size 4725048

layers.9.mlp/cfg.json

@@ -0,0 +1 @@
+{"expansion_factor": 32, "normalize_decoder": true, "num_latents": 768, "k": 32, "multi_topk": false, "skip_connection": false, "d_in": 768}

layers.9.mlp/sae.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9fee9934b042aa6269ae3df0a2b99ed29c7ddc50faf2df0ef898d3e6bcfec74f
+size 4725048

lr_scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46655426fcb2ea695eaf989d482d5a506c403f0d90388a363978fd1d7f2f81ce
+size 1268

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47ade93285e4ee2626dff3540fa65416a2e57e880ccf0170988c571acec7100a
+size 28947578

state.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89a0aeeb13376ffa3adb4ad040c1e02ac7ecb7591b9d0b95c4ad8d8ce6667e6d
+size 77138