Datasets:
Add eval config
Browse files
eval.yaml
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
name: News21InstructionRetrieval
|
| 2 |
+
description: Measuring retrieval instruction following ability on News21 narratives
|
| 3 |
+
for the FollowIR benchmark.
|
| 4 |
+
evaluation_framework: mteb
|
| 5 |
+
tasks:
|
| 6 |
+
- id: News21InstructionRetrieval
|
| 7 |
+
- id: News21InstructionRetrieval_default_test
|
| 8 |
+
config: default
|
| 9 |
+
split: test
|