JeongwonChoi
commited on
Commit
•
6d90905
1
Parent(s):
d6984dd
Update README.md
Browse files
README.md
CHANGED
@@ -59,22 +59,18 @@ text = """\
|
|
59 |
|
60 |
### **[Ko LM Eval Harnesss](https://github.com/Beomi/ko-lm-evaluation-harness)**
|
61 |
|
62 |
-
|
63 |
-
|
64 |
-
|
|
65 |
-
|
|
66 |
-
|
|
67 |
-
|
|
68 |
-
| kobest_hellaswag | 0.0 | 0.0 | 0.0 | 0.0 |
|
69 |
-
| kobest_sentineg | 0.0 | 0.0 | 0.0 | 0.0 |
|
70 |
|
71 |
### **[Ko-LLM-Leaderboard](https://huggingface.co/spaces/upstage/open-ko-llm-leaderboard)**
|
72 |
|
73 |
-
On Benchmarking ...
|
74 |
-
|
75 |
| Average | Ko-ARC | Ko-HellaSwag | Ko-MMLU | Ko-TruthfulQA | Ko-CommonGen V2 |
|
76 |
| ------: | -----: | -----------: | ------: | ------------: | --------------: |
|
77 |
-
|
|
78 |
|
79 |
## **Implementation Code**
|
80 |
|
|
|
59 |
|
60 |
### **[Ko LM Eval Harnesss](https://github.com/Beomi/ko-lm-evaluation-harness)**
|
61 |
|
62 |
+
| Task | 0-shot | 5-shot | 10-shot | 50-shot |
|
63 |
+
| :--------------- | -------: | -------: | -------: | -------: |
|
64 |
+
| kobest_boolq | 0.334282 | 0.516446 | 0.500478 | 0.498941 |
|
65 |
+
| kobest_copa | 0.515061 | 0.504321 | 0.492927 | 0.50809 |
|
66 |
+
| kobest_hellaswag | 0.36253 | 0.357733 | 0.355873 | 0.376502 |
|
67 |
+
| kobest_sentineg | 0.481146 | 0.657411 | 0.687417 | 0.635703 |
|
|
|
|
|
68 |
|
69 |
### **[Ko-LLM-Leaderboard](https://huggingface.co/spaces/upstage/open-ko-llm-leaderboard)**
|
70 |
|
|
|
|
|
71 |
| Average | Ko-ARC | Ko-HellaSwag | Ko-MMLU | Ko-TruthfulQA | Ko-CommonGen V2 |
|
72 |
| ------: | -----: | -----------: | ------: | ------------: | --------------: |
|
73 |
+
| 31.5 | 25.26 | 33.53 | 24.56 | 43.34 | 30.81 |
|
74 |
|
75 |
## **Implementation Code**
|
76 |
|