Rename msbench -> studybench in README
Browse files
README.md
CHANGED
|
@@ -20,9 +20,9 @@ configs:
|
|
| 20 |
data_files: openclaw.jsonl
|
| 21 |
---
|
| 22 |
|
| 23 |
-
#
|
| 24 |
|
| 25 |
-
**
|
| 26 |
open-source codebases, each paired with a **gold answer** and a **weighted, source-grounded grading
|
| 27 |
rubric**. The questions ask a model to produce working code that uses a specific library/framework
|
| 28 |
correctly; the rubric decomposes a correct answer into discrete, checkable claims, each tied to exact
|
|
@@ -38,8 +38,8 @@ Pick a subset with the second argument of `load_dataset`:
|
|
| 38 |
```python
|
| 39 |
from datasets import load_dataset
|
| 40 |
|
| 41 |
-
dspy = load_dataset("jacobli/
|
| 42 |
-
openclaw = load_dataset("jacobli/
|
| 43 |
```
|
| 44 |
|
| 45 |
| config | questions | topics | codebase |
|
|
|
|
| 20 |
data_files: openclaw.jsonl
|
| 21 |
---
|
| 22 |
|
| 23 |
+
# studybench
|
| 24 |
|
| 25 |
+
**studybench** is a small, high-effort benchmark of **expert-level coding questions** about real
|
| 26 |
open-source codebases, each paired with a **gold answer** and a **weighted, source-grounded grading
|
| 27 |
rubric**. The questions ask a model to produce working code that uses a specific library/framework
|
| 28 |
correctly; the rubric decomposes a correct answer into discrete, checkable claims, each tied to exact
|
|
|
|
| 38 |
```python
|
| 39 |
from datasets import load_dataset
|
| 40 |
|
| 41 |
+
dspy = load_dataset("jacobli/studybench", "dspy") # 30 questions
|
| 42 |
+
openclaw = load_dataset("jacobli/studybench", "openclaw") # 20 questions
|
| 43 |
```
|
| 44 |
|
| 45 |
| config | questions | topics | codebase |
|