Fix README: Goals are trajectory-specific, not constant

README.md

@@ -102,9 +102,13 @@ with torch.no_grad():
 
 ## Dataset Notes
 
-- All trajectories have the same start and end positions
-- Single demonstration repeated 70 times
-- Goal: Final cartesian position `[0.454, -0.133, 0.522]` (constant)
+- **Goals are trajectory-specific**: Each trajectory has a unique goal (final cartesian position)
+- **Goal distribution**: 
+  - X range: [-0.239, 0.294] meters
+  - Y range: [-0.284, 0.317] meters  
+  - Z range: [0.364, 0.579] meters
+- **Total trajectories**: ~500 trajectories across 20 H5 files
+- **Goal extraction**: Goal = final cartesian position of each trajectory (varies per trajectory)
 
 ## License