Update basic changes

.gitignore

@@ -0,0 +1,2 @@
+
+__pycache__

Makefile

@@ -0,0 +1,13 @@
+.PHONY: style format
+
+
+style:
+	python -m black --line-length 119 .
+	python -m isort .
+	ruff check --fix .
+
+
+quality:
+	python -m black --check --line-length 119 .
+	python -m isort --check-only .
+	ruff check .

app.py

@@ -1,5 +1,6 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
+from sae_auto_interp.sae import Sae
 
 """
 For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
@@ -40,24 +41,38 @@ def respond(
         yield response
 
 
+CITATION_BUTTON_TEXT = """
+@misc{zhang2024largemultimodalmodelsinterpret,
+      title={Large Multi-modal Models Can Interpret Features in Large Multi-modal Models},
+      author={Kaichen Zhang and Yifei Shen and Bo Li and Ziwei Liu},
+      year={2024},
+      eprint={2411.14982},
+      archivePrefix={arXiv},
+      primaryClass={cs.CV},
+      url={https://arxiv.org/abs/2411.14982},
+}
 """
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
-    ],
-)
+
+
+with gr.Blocks() as demo:
+    gr.Markdown(
+        """
+        # Large Multi-modal Models Can Interpret Features in Large Multi-modal Models
+
+        🔍 [ArXiv Paper](https://arxiv.org/abs/2411.14982) | 🏠 [LMMs-Lab Homepage](https://lmms-lab.framer.ai) | 🤗 [Huggingface Collections](https://huggingface.co/collections/lmms-lab/llava-sae-674026e4e7bc8c29c70bc3a3)
+        """
+    )
+
+    with gr.Tabs(elem_classes="tab-buttons") as tabs:
+        with gr.TabItem("Visualization of Activations", elem_id="visualization", id=0):
+            image = gr.Image()
+
+        with gr.TabItem("Steering Model", elem_id="steering", id=2):
+            chatbot = gr.Chatbot()
+
+    with gr.Row():
+        with gr.Accordion("📙 Citation", open=False):
+            gr.Markdown("```bib\n" + CITATION_BUTTON_TEXT + "\n```")
 
 
 if __name__ == "__main__":

requirements.txt

@@ -1 +1,3 @@
-huggingface_hub==0.25.2
+huggingface_hub==0.25.2
+gradio
+sae_auto_interp @ git+https://github.com/EvolvingLMMs-Lab/multimodal-sae