patrickNLP
commited on
Update README.md
Browse files
README.md
CHANGED
@@ -11,13 +11,13 @@ tags:
|
|
11 |
# BIRD-CRITIC-1.0-Flash
|
12 |
|
13 |
BIRD-Critic is the first SQL debugging benchmark designed to answer a critical question:
|
14 |
-
**Can large language models (LLMs)
|
15 |
Each task in BIRD-CRITIC has been verified by human experts on the following dimensions:
|
16 |
-
1) Reproduction of errors on BIRD env
|
17 |
2) Carefully curate test case functions for each task specifically.
|
18 |
- **Soft EX**: This metric can evaluate SELECT-ONLY tasks.
|
19 |
- **Soft EX + Parsing**: This metric can evaluate tasks with user specific requirements or refinements.
|
20 |
-
- **Test Case**: For CRUD (CREAT, READ, UPDATE,
|
21 |
- **Query Execution Plan**: For user tasks involving efficiency improvement or runtime errors, QEP can be introduced to evaluate solution SQLs on algorithm level.
|
22 |
4) Fast Eval Sandbox via PostgreSQL template & docker.
|
23 |
5) Created new RDBs in different scale and professional domains.
|
@@ -33,7 +33,7 @@ We are releasing a lite version of BIRD-Critic, `bird-critic-1.0-flash-exp`, whi
|
|
33 |
| Rank | Model Name | Score | Level |
|
34 |
|------|------------|-------|-----------|
|
35 |
| 1 | o1-preview-2024-09-12 | **38.5** | π Leading |
|
36 |
-
| 2 | deepseek-reasoner | 34.0 | π Elite |
|
37 |
| 3 | gpt-4o-2024-11-20 | 29.0 | π Elite |
|
38 |
| 4 | o1-mini | 28.0 | π Superior |
|
39 |
| 5 | deepseek-V3 | 27.5 | π Superior |
|
@@ -100,6 +100,7 @@ Below is a description of the dataset fields and additional information about th
|
|
100 |
- [ ] Release Full bird-critic-1.0-open (600 w/ 5 dialects).
|
101 |
- [ ] Release Full bird-critic-1.0-postgresql (600 pg tasks).
|
102 |
- [ ] Update agent baselines.
|
|
|
103 |
- [ ] BIRD-CRITIC 1.5 / 2.0 on track!
|
104 |
|
105 |
### Post By:
|
|
|
11 |
# BIRD-CRITIC-1.0-Flash
|
12 |
|
13 |
BIRD-Critic is the first SQL debugging benchmark designed to answer a critical question:
|
14 |
+
**Can large language models (LLMs) fix user issues in real-world database applications?** \
|
15 |
Each task in BIRD-CRITIC has been verified by human experts on the following dimensions:
|
16 |
+
1) Reproduction of errors on BIRD env to prevent data leakage.
|
17 |
2) Carefully curate test case functions for each task specifically.
|
18 |
- **Soft EX**: This metric can evaluate SELECT-ONLY tasks.
|
19 |
- **Soft EX + Parsing**: This metric can evaluate tasks with user specific requirements or refinements.
|
20 |
+
- **Test Case**: For DBA tasks, such as CRUD (CREAT, READ, UPDATE, DELET), test cases should be promised to evaluate the correct logic. This is also effective for user issues requiring multiple sequential SQLs to resolve.
|
21 |
- **Query Execution Plan**: For user tasks involving efficiency improvement or runtime errors, QEP can be introduced to evaluate solution SQLs on algorithm level.
|
22 |
4) Fast Eval Sandbox via PostgreSQL template & docker.
|
23 |
5) Created new RDBs in different scale and professional domains.
|
|
|
33 |
| Rank | Model Name | Score | Level |
|
34 |
|------|------------|-------|-----------|
|
35 |
| 1 | o1-preview-2024-09-12 | **38.5** | π Leading |
|
36 |
+
| 2 | deepseek-reasoner (r1) | 34.0 | π Elite |
|
37 |
| 3 | gpt-4o-2024-11-20 | 29.0 | π Elite |
|
38 |
| 4 | o1-mini | 28.0 | π Superior |
|
39 |
| 5 | deepseek-V3 | 27.5 | π Superior |
|
|
|
100 |
- [ ] Release Full bird-critic-1.0-open (600 w/ 5 dialects).
|
101 |
- [ ] Release Full bird-critic-1.0-postgresql (600 pg tasks).
|
102 |
- [ ] Update agent baselines.
|
103 |
+
- [ ] BIRD-Pro v0.5
|
104 |
- [ ] BIRD-CRITIC 1.5 / 2.0 on track!
|
105 |
|
106 |
### Post By:
|