Update dataset card task format description

README.md

@@ -19,7 +19,21 @@ tags:
 
 # RoboStressBench
 
-RoboStressBench is a robotics-oriented vision-language benchmark for evaluating model robustness under visual stress conditions. It contains multiple-choice questions, target grounding with bounding boxes, and placement/point grounding with ground-truth masks.
+RoboStressBench is a robotics-oriented vision-language benchmark for evaluating model robustness under visual stress conditions. It contains two task formats: visual question answering and localization tasks.
+
+## Task Formats
+
+### Visual Question Answering
+
+- Reasoning: interpret the scene under stress.
+- Planning: decide the next embodied action.
+- Spatial: answer relative-position and layout questions.
+- State Understanding: judge task completion or object state.
+
+### Localization Tasks
+
+- Placement Grounding: point to free space for placing an object.
+- Target Grounding: localize the target object with a bounding box or point.
 
 ## Dataset Structure