Spaces:
Sleeping
Sleeping
Update .gitattributes to streamline LFS tracking and modify README.md to reflect changes in project description and emoji
Browse files- .gitattributes +1 -15
- README.md +3 -4
.gitattributes
CHANGED
@@ -32,18 +32,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
32 |
*.xz filter=lfs diff=lfs merge=lfs -text
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
-
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
36 |
-
# Model Files (Examples - adjust as needed)
|
37 |
-
*.pt filter=lfs diff=lfs merge=lfs -text
|
38 |
-
*.pth filter=lfs diff=lfs merge=lfs -text
|
39 |
-
*.bin filter=lfs diff=lfs merge=lfs -text
|
40 |
-
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
41 |
-
*.onnx filter=lfs diff=lfs merge=lfs -text
|
42 |
-
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
43 |
-
|
44 |
-
# Data Files (Examples - uncomment/adjust if tracking large data)
|
45 |
-
# *.jsonl filter=lfs diff=lfs merge=lfs -text
|
46 |
-
# *.parquet filter=lfs diff=lfs merge=lfs -text
|
47 |
-
# *.arrow filter=lfs diff=lfs merge=lfs -text
|
48 |
-
# *.zip filter=lfs diff=lfs merge=lfs -text
|
49 |
-
# *.tar.gz filter=lfs diff=lfs merge=lfs -text
|
|
|
32 |
*.xz filter=lfs diff=lfs merge=lfs -text
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
README.md
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
---
|
2 |
title: PasswordLLM
|
3 |
-
emoji:
|
4 |
colorFrom: yellow
|
5 |
colorTo: gray
|
6 |
sdk: streamlit
|
@@ -10,7 +10,6 @@ pinned: false
|
|
10 |
short_description: Password-protect true intentions of an LLM to fake alignment
|
11 |
---
|
12 |
|
13 |
-
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
|
14 |
# Password-Protected LLM Intentions
|
15 |
|
16 |
This project explores the concept of creating an LLM that appears aligned under normal circumstances but reveals different intentions when prompted with a specific password.
|
@@ -26,8 +25,8 @@ This project explores the concept of creating an LLM that appears aligned under
|
|
26 |
## Structure
|
27 |
|
28 |
- `benchmarking/`: Contains all scripts, data, and results related to model evaluation.
|
29 |
-
- `finetuning/`: Contains all scripts
|
30 |
-
- `app.py`: A basic
|
31 |
- `requirements.txt`: Project dependencies.
|
32 |
- `.gitignore`: Standard git ignore configuration for Python/ML projects.
|
33 |
|
|
|
1 |
---
|
2 |
title: PasswordLLM
|
3 |
+
emoji: π
|
4 |
colorFrom: yellow
|
5 |
colorTo: gray
|
6 |
sdk: streamlit
|
|
|
10 |
short_description: Password-protect true intentions of an LLM to fake alignment
|
11 |
---
|
12 |
|
|
|
13 |
# Password-Protected LLM Intentions
|
14 |
|
15 |
This project explores the concept of creating an LLM that appears aligned under normal circumstances but reveals different intentions when prompted with a specific password.
|
|
|
25 |
## Structure
|
26 |
|
27 |
- `benchmarking/`: Contains all scripts, data, and results related to model evaluation.
|
28 |
+
- `finetuning/`: Contains all scripts and data related to fine-tuning.
|
29 |
+
- `app.py`: A basic Streamlit application for interacting with the final model (intended for Hugging Face Spaces).
|
30 |
- `requirements.txt`: Project dependencies.
|
31 |
- `.gitignore`: Standard git ignore configuration for Python/ML projects.
|
32 |
|