Datasets:
Upload eval.yaml with huggingface_hub
Browse files
eval.yaml
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
name: YC-Bench
|
| 2 |
+
description: >
|
| 3 |
+
Long-horizon agent benchmark where an LLM plays CEO of an AI startup
|
| 4 |
+
over a simulated 1-year run via CLI tool use. Tests sustained strategic
|
| 5 |
+
coherence: resource management, deadline reasoning, prestige climbing,
|
| 6 |
+
and adversarial client detection.
|
| 7 |
+
evaluation_framework: yc-bench
|
| 8 |
+
|
| 9 |
+
tasks:
|
| 10 |
+
- id: medium
|