Datasets:
Remove unrelated EVIDENT evaluation links from Ko-Bench
Browse files
README.md
CHANGED
|
@@ -29,8 +29,6 @@ configs:
|
|
| 29 |
|
| 30 |
# Ko-Bench: Korean Web GUI Evaluation
|
| 31 |
|
| 32 |
-
[](https://github.com/yerinNam/EVIDENT)
|
| 33 |
-
|
| 34 |
**Ko-Bench was created by human annotators.**
|
| 35 |
|
| 36 |
The suite brings together three Korean benchmarks for GUI grounding, offline
|
|
@@ -113,12 +111,6 @@ Preview selection is illustrative and is not intended for reporting benchmark sc
|
|
| 113 |
Pixel coordinates and normalized `*_norm` fields retain their recorded
|
| 114 |
conventions. No action vocabulary or coordinate conversion is applied in this release.
|
| 115 |
|
| 116 |
-
For evidence-grounded evaluation of collected GUI/web trajectories, see
|
| 117 |
-
[yerinNam/EVIDENT](https://github.com/yerinNam/EVIDENT). EVIDENT extracts
|
| 118 |
-
screenshot evidence, reviews final-answer claims, and applies task rubrics;
|
| 119 |
-
its task metadata and judge pipeline are documented in that repository.
|
| 120 |
-
It is a related trajectory judge, rather than the offline grounding scorer.
|
| 121 |
-
|
| 122 |
## About KORMo-VL
|
| 123 |
|
| 124 |
KORMo-VL is a vision-language model developed from scratch by the
|
|
|
|
| 29 |
|
| 30 |
# Ko-Bench: Korean Web GUI Evaluation
|
| 31 |
|
|
|
|
|
|
|
| 32 |
**Ko-Bench was created by human annotators.**
|
| 33 |
|
| 34 |
The suite brings together three Korean benchmarks for GUI grounding, offline
|
|
|
|
| 111 |
Pixel coordinates and normalized `*_norm` fields retain their recorded
|
| 112 |
conventions. No action vocabulary or coordinate conversion is applied in this release.
|
| 113 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 114 |
## About KORMo-VL
|
| 115 |
|
| 116 |
KORMo-VL is a vision-language model developed from scratch by the
|