init
Browse files
README.md
CHANGED
|
@@ -33,14 +33,16 @@ This dataset contains 5 different word analogy questions used in [Analogy Langua
|
|
| 33 |
### Data Instances
|
| 34 |
An example of `test` looks as follows.
|
| 35 |
```
|
| 36 |
-
{
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
|
|
|
|
|
|
| 42 |
```
|
| 43 |
-
|
| 44 |
and `answer` indicates the index of correct candidate which starts from `0`.
|
| 45 |
All data is lowercased except Google dataset.
|
| 46 |
|
|
|
|
| 33 |
### Data Instances
|
| 34 |
An example of `test` looks as follows.
|
| 35 |
```
|
| 36 |
+
{
|
| 37 |
+
"stem": ["raphael", "painter"],
|
| 38 |
+
"answer": 2,
|
| 39 |
+
"choice": [["andersen", "plato"],
|
| 40 |
+
["reading", "berkshire"],
|
| 41 |
+
["marx", "philosopher"],
|
| 42 |
+
["tolstoi", "edison"]]
|
| 43 |
+
}
|
| 44 |
```
|
| 45 |
+
The `stem` is the query word pair, `choice` has word pair candidates,
|
| 46 |
and `answer` indicates the index of correct candidate which starts from `0`.
|
| 47 |
All data is lowercased except Google dataset.
|
| 48 |
|