Update README.md

README.md

@@ -12,9 +12,11 @@ dataset_info:
   - name: prompt
     dtype: string
   - name: images_by_turn
-    dtype: string
+    sequence:
+      sequence:
+        dtype: image
   - name: rubrics
-    dtype: string
+    sequence: string
   splits:
   - name: train
     num_bytes: 3506927656
@@ -28,7 +30,7 @@ dataset_info:
       path: data/*.parquet
 ---
 
-VisuAlToolBench is a challenging benchmark to assess tool-enabled visual perception, transformation, and reasoning in multimodal LLMs. It evaluates whether models can not only think about images but also think with images by actively manipulating visuals (e.g., crop, edit, enhance) and integrating general-purpose tools to solve complex tasks. The dataset contains single-turn and multi-turn tasks across diverse domains, each accompanied by detailed rubrics for systematic evaluation.
+VisuAlToolBench is a challenging benchmark to assess tool-enabled visual perception, transformation, and reasoning in multimodal LLMs. It evaluates whether models can not only think about images but also think with images by actively manipulating visuals (e.g., crop, edit, enhance) and integrating general-purpose tools to solve complex tasks. The dataset contains single-turn and multi-turn tasks across diverse domains, each accompanied by detailed rubrics for systematic evaluation. Parquet files under `data/` are auto-indexed by the Hub and power the Dataset Viewer.
 
 Paper: [BEYOND SEEING: Evaluating Multimodal LLMs on Tool-enabled Image Perception, Transformation, and Reasoning](https://static.scale.com/uploads/654197dc94d34f66c0f5184e/vtb_paper.pdf)