Create eval.yaml
Browse files
eval.yaml
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
name: WildClawBench
|
| 2 |
+
description: Hard, practical, end-to-end evaluation for AI agents — in the wild.
|
| 3 |
+
evaluation_framework: inspect-ai
|
| 4 |
+
tasks:
|
| 5 |
+
- id: overall
|
| 6 |
+
name: Overall
|
| 7 |
+
config: default
|
| 8 |
+
split: test
|