Upload folder using huggingface_hub

README.md

@@ -0,0 +1,67 @@
+---
+library_name: sae_lens
+tags:
+  - sparse-autoencoder
+  - mechanistic-interpretability
+  - sae
+---
+
+# Sparse Autoencoders for Qwen/Qwen2.5-7B-Instruct
+
+This repository contains 3 Sparse Autoencoder(s) (SAE) trained using [SAELens](https://github.com/jbloomAus/SAELens).
+
+## Model Details
+
+| Property | Value |
+|----------|-------|
+| **Base Model** | `Qwen/Qwen2.5-7B-Instruct` |
+| **Architecture** | `gated` |
+| **Input Dimension** | 3584 |
+| **SAE Dimension** | 16384 |
+| **Training Dataset** | `TQRG/DeltaSecommits_qwen-2.5-7b-instruct_tokenized` |
+
+## Available Hook Points
+
+| Hook Point |
+|------------|
+| `blocks.0.hook_resid_post` |
+| `blocks.14.hook_resid_post` |
+| `blocks.27.hook_resid_post` |
+
+## Usage
+
+```python
+from sae_lens import SAE
+
+# Load an SAE for a specific hook point
+sae, cfg_dict, sparsity = SAE.from_pretrained(
+    release="rufimelo/secure_code_qwen_coder_gated_16384",
+    sae_id="blocks.0.hook_resid_post"  # Choose from available hook points above
+)
+
+# Use with TransformerLens
+from transformer_lens import HookedTransformer
+
+model = HookedTransformer.from_pretrained("Qwen/Qwen2.5-7B-Instruct")
+
+# Get activations and encode
+_, cache = model.run_with_cache("your text here")
+activations = cache["blocks.0.hook_resid_post"]
+features = sae.encode(activations)
+```
+
+## Files
+
+- `blocks.0.hook_resid_post/cfg.json` - SAE configuration
+- `blocks.0.hook_resid_post/sae_weights.safetensors` - Model weights
+- `blocks.0.hook_resid_post/sparsity.safetensors` - Feature sparsity statistics
+- `blocks.14.hook_resid_post/cfg.json` - SAE configuration
+- `blocks.14.hook_resid_post/sae_weights.safetensors` - Model weights
+- `blocks.14.hook_resid_post/sparsity.safetensors` - Feature sparsity statistics
+- `blocks.27.hook_resid_post/cfg.json` - SAE configuration
+- `blocks.27.hook_resid_post/sae_weights.safetensors` - Model weights
+- `blocks.27.hook_resid_post/sparsity.safetensors` - Feature sparsity statistics
+
+## Training
+
+These SAEs were trained with SAELens version 6.26.2.

blocks.0.hook_resid_post/cfg.json

@@ -0,0 +1 @@
+{"metadata": {"sae_lens_version": "6.26.2", "sae_lens_training_version": "6.26.2", "dataset_path": "TQRG/DeltaSecommits_qwen-2.5-7b-instruct_tokenized", "hook_name": "blocks.0.hook_resid_post", "model_name": "Qwen/Qwen2.5-7B-Instruct", "model_class_name": "HookedTransformer", "hook_head_index": null, "context_size": 128, "seqpos_slice": [null, null], "model_from_pretrained_kwargs": {}, "prepend_bos": true, "exclude_special_tokens": false, "sequence_separator_token": "bos", "disable_concat_sequences": false}, "apply_b_dec_to_input": true, "d_sae": 16384, "normalize_activations": "layer_norm", "d_in": 3584, "reshape_activations": "none", "dtype": "float32", "device": "cuda", "architecture": "gated"}

blocks.0.hook_resid_post/sae_weights.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f6c80a1e739e28e4557214f520a22c0c5714145d22050f75f35de66283a2bb1
+size 469973472

blocks.0.hook_resid_post/sparsity.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dc69d6855f9478251df097b610c5972ac4dedaa897eadc549fc8579e6a02d4e9
+size 65616

blocks.14.hook_resid_post/cfg.json

@@ -0,0 +1 @@
+{"device": "cuda", "dtype": "float32", "metadata": {"sae_lens_version": "6.26.2", "sae_lens_training_version": "6.26.2", "dataset_path": "TQRG/DeltaSecommits_qwen-2.5-7b-instruct_tokenized", "hook_name": "blocks.14.hook_resid_post", "model_name": "Qwen/Qwen2.5-7B-Instruct", "model_class_name": "HookedTransformer", "hook_head_index": null, "context_size": 128, "seqpos_slice": [null, null], "model_from_pretrained_kwargs": {}, "prepend_bos": true, "exclude_special_tokens": false, "sequence_separator_token": "bos", "disable_concat_sequences": false}, "d_sae": 16384, "d_in": 3584, "apply_b_dec_to_input": true, "reshape_activations": "none", "normalize_activations": "layer_norm", "architecture": "gated"}

blocks.14.hook_resid_post/sae_weights.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7c303fab8a318e69a408cdeb5268a9b0d7d5407e7fb1940cbef810b82ddc732
+size 469973472

blocks.14.hook_resid_post/sparsity.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:41cbdd824edbc837016cd6a0a9b88d2f1500543750dc6814d778f358fdeed1a1
+size 65616

blocks.27.hook_resid_post/cfg.json

@@ -0,0 +1 @@
+{"d_sae": 16384, "device": "cuda", "normalize_activations": "layer_norm", "metadata": {"sae_lens_version": "6.26.2", "sae_lens_training_version": "6.26.2", "dataset_path": "TQRG/DeltaSecommits_qwen-2.5-7b-instruct_tokenized", "hook_name": "blocks.27.hook_resid_post", "model_name": "Qwen/Qwen2.5-7B-Instruct", "model_class_name": "HookedTransformer", "hook_head_index": null, "context_size": 128, "seqpos_slice": [null, null], "model_from_pretrained_kwargs": {}, "prepend_bos": true, "exclude_special_tokens": false, "sequence_separator_token": "bos", "disable_concat_sequences": false}, "reshape_activations": "none", "d_in": 3584, "dtype": "float32", "apply_b_dec_to_input": true, "architecture": "gated"}

blocks.27.hook_resid_post/sae_weights.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52bdfe19b475f7bf38ff376237a7dc9c189a3571b300dad46af14ea2507d0e7c
+size 469973472

blocks.27.hook_resid_post/sparsity.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e04d3c8b8af063910438e1bc402af51ce983cbb4eb8c82a63e8bc6ba425b5313
+size 65616