Upload 3 files
Browse filesChanges in this version:
1. removed some templates
2. fixed all object names to be singular
3. filtered names that are tokenized to a single token on GPT-2, Mistral, Phi, Llama, Qwen and Pythia models.
4. 10k samples in each split
5. added abc counterfactuals
- .gitattributes +3 -0
- dev.json +0 -0
- test.json +0 -0
- train.json +0 -0
.gitattributes
CHANGED
|
@@ -53,3 +53,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 53 |
*.jpg filter=lfs diff=lfs merge=lfs -text
|
| 54 |
*.jpeg filter=lfs diff=lfs merge=lfs -text
|
| 55 |
*.webp filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
*.jpg filter=lfs diff=lfs merge=lfs -text
|
| 54 |
*.jpeg filter=lfs diff=lfs merge=lfs -text
|
| 55 |
*.webp filter=lfs diff=lfs merge=lfs -text
|
| 56 |
+
dev.json filter=lfs diff=lfs merge=lfs -text
|
| 57 |
+
test.json filter=lfs diff=lfs merge=lfs -text
|
| 58 |
+
train.json filter=lfs diff=lfs merge=lfs -text
|
dev.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
test.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
train.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|