Create split BFCL v4 dataset
Browse files- README.md +48 -0
- eval.jsonl +0 -0
- train.jsonl +0 -0
README.md
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
task_categories:
|
| 6 |
+
- text-generation
|
| 7 |
+
pretty_name: BFCL V4 Parallel Native
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
# BFCL V4 Parallel Native
|
| 11 |
+
|
| 12 |
+
Native BFCL v4 single-turn parallel function-calling rows for decentralized multi-agent collaboration.
|
| 13 |
+
|
| 14 |
+
Source data comes from the official Berkeley Function Calling Leaderboard v4 data and possible-answer files.
|
| 15 |
+
|
| 16 |
+
## Fields
|
| 17 |
+
|
| 18 |
+
- `id`
|
| 19 |
+
- `official_category`
|
| 20 |
+
- `task_type`
|
| 21 |
+
- `user_prompt`
|
| 22 |
+
- `function`
|
| 23 |
+
- `ground_truth`
|
| 24 |
+
|
| 25 |
+
## Categories
|
| 26 |
+
|
| 27 |
+
- `live_parallel`
|
| 28 |
+
- `live_parallel_multiple`
|
| 29 |
+
- `parallel`
|
| 30 |
+
- `parallel_multiple`
|
| 31 |
+
|
| 32 |
+
## Counts
|
| 33 |
+
|
| 34 |
+
- `train`: 352 rows
|
| 35 |
+
- `eval`: 88 rows
|
| 36 |
+
- total: 440 rows
|
| 37 |
+
|
| 38 |
+
By category:
|
| 39 |
+
|
| 40 |
+
- `live_parallel`: 16
|
| 41 |
+
- `live_parallel_multiple`: 24
|
| 42 |
+
- `parallel`: 200
|
| 43 |
+
- `parallel_multiple`: 200
|
| 44 |
+
|
| 45 |
+
## Notes
|
| 46 |
+
|
| 47 |
+
- Rows are native single-turn BFCL parallel or live parallel examples.
|
| 48 |
+
- No `turn_index` field is included because these rows are not flattened multi-turn examples.
|
eval.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
train.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|