Add README with per-config metadata

README.md

@@ -2,19 +2,19 @@
 license: cc-by-4.0
 pretty_name: COSPLAN
 configs:
-- config_name: checkboard
+- config_name: Maze-E
   data_files:
   - split: test
     path: checkboard/metadata.jsonl
-- config_name: blocksworld
+- config_name: Blocksworld-E
   data_files:
   - split: test
     path: blocksworld/metadata.jsonl
-- config_name: imagenet
+- config_name: Shuffle-E
   data_files:
   - split: test
     path: imagenet/metadata.jsonl
-- config_name: robovqa
+- config_name: Robovqa-E
   data_files:
   - split: test
     path: robovqa/metadata.jsonl
@@ -32,17 +32,17 @@ erroneous step) and an image, select the option that correctly completes the tas
 
 | Config | Items | Task |
 |--------|-------|------|
-| `checkboard`   | 5000 | Complete the path from start to goal around obstacles |
-| `blocksworld`  | 5000 | Complete the block-move sequence to the target configuration |
-| `imagenet`     | 5000 | Complete the patch-swap sequence to rearrange into the target image |
-| `robovqa`      |  390 | Complete the robot action sequence |
+| `Maze-E`         | 5000 | Complete the path from start to goal around obstacles |
+| `Blocksworld-E`  | 5000 | Complete the block-move sequence to the target configuration |
+| `Shuffle-E`      | 5000 | Complete the patch-swap sequence to rearrange into the target image |
+| `Robovqa-E`      |  390 | Complete the robot action sequence |
 
 ## Usage
 
 ```python
 from datasets import load_dataset
 
-ds = load_dataset("shrg7/COSPLAN", "blocksworld", split="test")
+ds = load_dataset("shrg7/COSPLAN", "Blocksworld-E", split="test")
 
 row = ds[0]
 print(row["prompt"], row["correct_option"])