Upload folder using huggingface_hub

DeepSeek-R1-Distill-LLama-8B/s2_start_2_n_end_2/README.md

@@ -0,0 +1,9 @@
+## Scenario 2: Positive at `<think> + 2` and Negative at `</think> +  2` (The first disntint tokens of thinking and answer parts)
+
+- Sources of activations: [DeepSeek-R1-Distill-LLama-8B/activations](https://huggingface.co/datasets/LoneResearch/thinking-model-activations/tree/main/DeepSeek-R1-Distill-LLama-8B/activations)
+
+```python
+BATCH_SIZE = 128
+CHOSEN_EXTRACTION_POINT = 45
+RANDOM_SEED = 42
+```

DeepSeek-R1-Distill-LLama-8B/s2_start_2_n_end_2/chosen_steering_vector.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:023c8bb71c751deb73f90b85c27e4b4b22c989585d87003bbf0529d2e7d0af92
+size 525607

DeepSeek-R1-Distill-LLama-8B/s2_start_2_n_end_2/chosen_steering_vector_normed.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d9272862f4f65888b8777d2ee3d88669fa912b0d697d62244332e084b2e2b71
+size 9546