Add dataset card

README.md

@@ -27,6 +27,12 @@ dataset_info:
     dtype: int32
   - name: question_type
     dtype: string
+  - name: question
+    dtype: string
+  - name: notation
+    dtype: string
+  - name: notation_type
+    dtype: string
   - name: option_a
     dtype: string
   - name: option_b
@@ -39,65 +45,59 @@ dataset_info:
     dtype: string
   - name: correct_idx
     dtype: int32
-  - name: notation
-    dtype: string
-  - name: notation_type
-    dtype: string
-  - name: question
-    dtype: string
   - name: image
     dtype: image
   splits:
   - name: fork
-    num_bytes: 76992.0
+    num_bytes: 0
     num_examples: 200
   - name: legal
-    num_bytes: 43330.0
+    num_bytes: 0
     num_examples: 200
   - name: puzzle
-    num_bytes: 49953.0
+    num_bytes: 0
     num_examples: 200
   - name: eval
-    num_bytes: 106305.0
+    num_bytes: 0
     num_examples: 200
   - name: carbon
-    num_bytes: 45878.0
+    num_bytes: 0
     num_examples: 200
   - name: hydrogen
-    num_bytes: 46575.0
+    num_bytes: 0
     num_examples: 200
   - name: weight
-    num_bytes: 45289.0
+    num_bytes: 0
     num_examples: 200
   - name: caption
-    num_bytes: 341320.0
+    num_bytes: 0
     num_examples: 200
   - name: notes
-    num_bytes: 141484.0
+    num_bytes: 0
     num_examples: 200
   - name: measures
-    num_bytes: 128298.0
+    num_bytes: 0
     num_examples: 200
   - name: forms
-    num_bytes: 130491.0
+    num_bytes: 0
     num_examples: 200
   - name: rhythm
-    num_bytes: 115494.0
+    num_bytes: 0
     num_examples: 200
   - name: path_counting
-    num_bytes: 76764.0
+    num_bytes: 0
     num_examples: 200
   - name: path_existence
-    num_bytes: 85765.0
+    num_bytes: 0
     num_examples: 200
   - name: shortest_path
-    num_bytes: 75974.0
+    num_bytes: 0
     num_examples: 200
   - name: bfs_traversal
-    num_bytes: 92605.0
+    num_bytes: 0
     num_examples: 200
-  download_size: 654451
-  dataset_size: 1602517.0
+  download_size: 0
+  dataset_size: 0
 configs:
 - config_name: default
   data_files:
@@ -155,8 +155,8 @@ Evaluating whether vision–language models (VLMs) reason consistently across re
 - **4 Domains**: Chess, Chemistry, Music, Graph Theory with standardized notations
 - **16 Tasks**: 4 tasks per domain (64 total task-modality combinations)
 - **3 Modalities**: Language-only (L), Vision-only (V), Vision-Language (VL)
-- **3,200 Questions**: 200 questions × 16 tasks
-- **9,600 Evaluations**: Each question evaluated in all 3 modalities
+- **3,200 Base Samples**: 200 samples × 16 tasks
+- **9,600 Evaluations**: TaskLoader generates 3 modality-specific prompts per base sample
 - **Semantic Equivalence**: Same information presented in different representational formats
 
 ## Domains and Notation Systems
@@ -190,7 +190,7 @@ The dataset is organized into 16 task-based splits (600 samples each):
 - **Music**: `notes`, `measures`, `forms`, `rhythm`
 - **Graph Theory**: `path_counting`, `path_existence`, `shortest_path`, `bfs_traversal`
 
-Each split contains 600 samples: 200 base questions × 3 modalities (L, V, VL).
+Each split contains 200 base samples. TaskLoader generates modality-specific prompts (L, V, VL) from these base samples.
 
 ## Usage
 
@@ -204,16 +204,15 @@ dataset = load_dataset("lilvjosephtang/SEAM-Benchmark")
 chess_fork = dataset["fork"]  # Chess fork detection (600 samples)
 chemistry_carbon = dataset["carbon"]  # Carbon atom counting (600 samples)
 
-# Filter by modality within a task
-language_only = chess_fork.filter(lambda x: x["modality"] == "Language")
-vision_only = chess_fork.filter(lambda x: x["modality"] == "Vision") 
-vision_language = chess_fork.filter(lambda x: x["modality"] == "Vision-Language")
+# Each task contains 200 base samples
+# TaskLoader generates modality-specific prompts (L/V/VL) from these base samples
+print(f"Task {chess_fork[0]['task']} has {len(chess_fork)} base samples")
 
 # Example sample structure
 sample = chess_fork[0]
 print(f"Task: {sample['task']}")
 print(f"Domain: {sample['domain']}")
-print(f"Modality: {sample['modality']}")
+# No modality field - TaskLoader handles modality generation
 print(f"Question: {sample['question']}")
 print(f"Options: A) {sample['option_a']}, B) {sample['option_b']}, C) {sample['option_c']}, D) {sample['option_d']}")
 print(f"Correct Answer: {sample['correct_answer']}")
@@ -226,7 +225,7 @@ print(f"Notation: {sample['notation']}")  # FEN string for chess
 Each sample contains:
 - `task`: Task identifier (e.g., "fork", "carbon")
 - `domain`: Domain category ("chess", "chemistry", "music", "graph")
-- `modality`: Evaluation modality ("Language", "Vision", "Vision-Language")
+- No modality field (TaskLoader generates modality-specific prompts)
 - `index`: Sample index within the task
 - `question`: Question text (if applicable)
 - `notation`: Domain-specific notation (FEN, SMILES, ABC, adjacency matrix)
@@ -234,7 +233,7 @@ Each sample contains:
 - `option_a`, `option_b`, `option_c`, `option_d`: Multiple choice options
 - `correct_answer`: The correct answer
 - `correct_idx`: Index of the correct option
-- `image`: Associated image (PIL Image for Vision/Vision-Language modalities, None for Language)
+- `image`: Associated image (PIL Image, None for base storage - TaskLoader handles image loading for V/VL modalities)
 
 ## Evaluation Protocol