diff --git a/.opencode/agents/paper-image-extractor.md b/.opencode/agents/paper-image-extractor.md
index e781c7d..fa3362f 100644
--- a/.opencode/agents/paper-image-extractor.md
+++ b/.opencode/agents/paper-image-extractor.md
@@ -32,12 +32,29 @@ matches = re.findall(pattern, paper_content)
 # Returns: [(alt_text, image_path), ...]
 ```
 
-### Step 2: Analyze Each Image
+### Step 2: Read and Analyze Each Image
+
+**CRITICAL**: You MUST use the `read` tool on each image file to visually analyze it.
 
 For each image found:
-1. Read the image file
-2. Analyze with vision capabilities
-3. Generate corresponding Python plotting code
+1. **Use the `read` tool on the image file path** - This returns the image for visual analysis
+2. Analyze what you **SEE** in the image (not what the paper text says about it)
+3. Extract precise data points, colors, line styles, axis ranges from the actual image
+4. Generate corresponding Python plotting code based on your visual analysis
+
+**Example workflow:**
+```
+# First, use read tool on the image
+read(filePath="path/to/figure1.png")  
+
+# Then analyze what you SEE:
+# - How many curves/bars/elements?
+# - What are the axis labels and ranges?
+# - What are the approximate data values at key points?
+# - What colors and line styles are used?
+```
+
+**DO NOT** rely solely on text descriptions in the paper. The paper text may be incomplete or ambiguous. Your understanding must come from **SEEING** the actual images.
 
 ### Step 3: Generate Outputs
 
@@ -45,6 +62,39 @@ Create two outputs in `analysis/` directory:
 1. `image_understanding.md` - Brief descriptions
 2. `reference_plots.py` - Self-contained plotting script
 
+### Step 4: Verify Your Understanding
+
+After generating `reference_plots.py`:
+1. Run the script: `python analysis/reference_plots.py`
+2. Open and compare your generated images with the originals
+3. If they don't match (wrong chart type, missing curves, wrong trends), **re-read the original images** and fix your code
+4. Repeat until your reproductions capture the essential structure
+
+## Extracting Data from Images
+
+When you read an image file with the `read` tool, you see it visually. Extract data by:
+
+### For Line Plots
+- Count the number of curves and identify each by color/style
+- Estimate Y values at regular X intervals (e.g., every 10 units)
+- Note the axis ranges and labels
+- Use `scipy.interpolate.PchipInterpolator` for smooth curves from sparse points
+
+### For Bar Charts
+- Read the exact bar heights from the Y-axis
+- Note category labels on X-axis
+- Count number of groups and bars per group
+
+### For Architecture Diagrams
+- List all components/blocks
+- Note the connections and data flow direction
+- Extract any dimension annotations (e.g., "B×T×D")
+
+### For Scatter Plots
+- Estimate cluster centers and spread
+- Note any trend lines or boundaries
+- Identify different marker types/colors
+
 ## Required Outputs
 
 ### 1. image_understanding.md
@@ -163,33 +213,40 @@ if __name__ == "__main__":
 
 ## Guidelines for Plot Generation
 
+**Key Principle**: Extract data from what you SEE in the image, not from paper text.
+
 ### For Training Curves
-- Extract approximate data points from the image
-- Use numpy to generate smooth curves matching the trend
+- Read the image first, count the curves, identify colors
+- Extract approximate data points at regular intervals from the image
+- Use `scipy.interpolate.PchipInterpolator` for smooth interpolation
 - Include axis labels matching the paper
 
 ### For Architecture Diagrams
 - Create simplified block diagrams showing data flow
-- Label input/output shapes
+- Label input/output shapes as seen in the figure
 - Show key components (attention, FFN, etc.)
 
 ### For Bar Charts / Tables
-- Extract the numerical values
+- Extract the numerical values by reading from the axis in the image
 - Recreate using matplotlib bar plots
+- Match the grouping and colors
 
 ### For Scatter Plots / Comparisons
-- Approximate the data distribution
+- Estimate data point positions from the image
 - Maintain relative positions and trends
+- Match marker styles and colors
 
 ## Important Notes
 
-1. **Minimal prompting**: When analyzing images, let the multimodal model understand naturally. Avoid over-specifying what to look for.
+1. **READ THE IMAGES**: Use the `read` tool on every image file. Do not skip this step. Your analysis quality depends on actually seeing the images.
 
-2. **Approximate is OK**: The goal is to verify understanding, not pixel-perfect reproduction. Trends and key values matter more than exact matches.
+2. **Visual over textual**: If the paper text says "Figure 3 shows X" but you see Y in the image, trust what you SEE.
 
-3. **Self-contained script**: The reference_plots.py must run without external dependencies beyond numpy/matplotlib.
+3. **Approximate is OK**: The goal is to verify understanding, not pixel-perfect reproduction. Trends and key values matter more than exact matches.
 
-4. **Data source labels**: Always note in comments that values are "extracted from paper figure" - this flags them as reference only, not ground truth.
+4. **Self-contained script**: The reference_plots.py must run without external dependencies beyond numpy/matplotlib/scipy.
+
+5. **Data source labels**: Always note in comments that values are "extracted from paper figure" - this flags them as reference only, not ground truth.
 
 ## Quality Checklist