Update README.md
Browse files
README.md
CHANGED
@@ -39,6 +39,19 @@ LLaMA-Pro-Ko's performance is evaluated on two fronts: its proficiency in Englis
|
|
39 |
|
40 |
### Korean Evaluation
|
41 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
42 |
#### KoBEST
|
43 |
|
44 |
**5shot**
|
@@ -61,14 +74,6 @@ LLaMA-Pro-Ko's performance is evaluated on two fronts: its proficiency in Englis
|
|
61 |
|
62 |
|
63 |
|
64 |
-
#### Open Ko LLM Benchmark
|
65 |
-
|
66 |
-
| | Ko-ARC | Ko-HellaSwag | Ko-MMLU | Ko-TruthfulQA | Ko-CommonGen V2 | AVG |
|
67 |
-
| ------------------------------------------------------------ | --------- | ------------ | --------- | ------------- | --------------- | --------- |
|
68 |
-
| [Llama-2-7b](https://huggingface.co/NousResearch/Nous-Hermes-llama-2-7b) | 31.91 | 41.68 | 34.11 | 48.49 | 30.34 | 37.31 |
|
69 |
-
| [beomi/open-llama-2-ko-7b](https://huggingface.co/beomi/open-llama-2-ko-7b) | 40.02 | 50.27 | 27.60 | 38.67 | 42.15 | 39.74 |
|
70 |
-
| llama-pro-ko-8b | **40.19** | **51.26** | **36.80** | **40.24** | **43.8** | **42.46** |
|
71 |
-
|
72 |
|
73 |
|
74 |
### English Evaluation
|
|
|
39 |
|
40 |
### Korean Evaluation
|
41 |
|
42 |
+
|
43 |
+
#### Open Ko LLM Benchmark
|
44 |
+
|
45 |
+
| | Ko-ARC | Ko-HellaSwag | Ko-MMLU | Ko-TruthfulQA | Ko-CommonGen V2 | AVG |
|
46 |
+
| ------------------------------------------------------------ | --------- | ------------ | --------- | ------------- | --------------- | --------- |
|
47 |
+
| [Llama-2-7b](https://huggingface.co/NousResearch/Nous-Hermes-llama-2-7b) | 31.91 | 41.68 | 34.11 | 48.49 | 30.34 | 37.31 |
|
48 |
+
| [beomi/open-llama-2-ko-7b](https://huggingface.co/beomi/open-llama-2-ko-7b) | 40.02 | 50.27 | 27.60 | 38.67 | 42.15 | 39.74 |
|
49 |
+
| llama-pro-ko-8b | **40.19** | **51.26** | **36.80** | **40.24** | **43.8** | **42.46** |
|
50 |
+
|
51 |
+
|
52 |
+
|
53 |
+
|
54 |
+
|
55 |
#### KoBEST
|
56 |
|
57 |
**5shot**
|
|
|
74 |
|
75 |
|
76 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
77 |
|
78 |
|
79 |
### English Evaluation
|