Update README.md
Browse files
README.md
CHANGED
@@ -29,27 +29,6 @@ The following models were included in the merge:
|
|
29 |
|
30 |
### Score
|
31 |
```
|
32 |
-
// use llm-ko-eval
|
33 |
-
"scores": {
|
34 |
-
"AVG_llm_kr_eval": "0.4425",
|
35 |
-
"EL": "0.0522",
|
36 |
-
"FA": "0.0865",
|
37 |
-
"NLI": "0.6700",
|
38 |
-
"QA": "0.5100",
|
39 |
-
"RC": "0.8937",
|
40 |
-
"klue_ner_set_f1": "0.0944",
|
41 |
-
"klue_re_exact_match": "0.0100",
|
42 |
-
"kmmlu_preview_exact_match": "0.4000",
|
43 |
-
"kobest_copa_exact_match": "0.8200",
|
44 |
-
"kobest_hs_exact_match": "0.5500",
|
45 |
-
"kobest_sn_exact_match": "0.9800",
|
46 |
-
"kobest_wic_exact_match": "0.6200",
|
47 |
-
"korea_cg_bleu": "0.0865",
|
48 |
-
"kornli_exact_match": "0.6400",
|
49 |
-
"korsts_pearson": "0.8547",
|
50 |
-
"korsts_spearman": "0.8464"
|
51 |
-
}
|
52 |
-
|
53 |
openai/gpt-4 : 0.6158
|
54 |
gemini-pro: 0.515
|
55 |
OpenCarrot-Mix-7B (this) : 0.4425
|
@@ -57,19 +36,40 @@ mistralai/Mixtral-8x7B-Instruct-v0.1 : 0.4304
|
|
57 |
openai/gpt-3.5-turbo : 0.4217
|
58 |
```
|
59 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
60 |
|
61 |
-
|
62 |
-
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
```
|
72 |
|
|
|
|
|
|
|
73 |
|
74 |
### Configuration
|
75 |
|
|
|
29 |
|
30 |
### Score
|
31 |
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
32 |
openai/gpt-4 : 0.6158
|
33 |
gemini-pro: 0.515
|
34 |
OpenCarrot-Mix-7B (this) : 0.4425
|
|
|
36 |
openai/gpt-3.5-turbo : 0.4217
|
37 |
```
|
38 |
|
39 |
+
| ํ๊ฐ ์งํ | ์ ์ |
|
40 |
+
|--------------|---------|
|
41 |
+
| AVG_llm_kr_eval | 0.4425 |
|
42 |
+
| EL | 0.0522 |
|
43 |
+
| FA | 0.0865 |
|
44 |
+
| NLI | 0.6700 |
|
45 |
+
| QA | 0.5100 |
|
46 |
+
| RC | 0.8937 |
|
47 |
+
| klue_ner_set_f1| 0.0944 |
|
48 |
+
| klue_re_exact_match | 0.0100 |
|
49 |
+
| kmmlu_preview_exact_match | 0.4000 |
|
50 |
+
| kobest_copa_exact_match | 0.8200 |
|
51 |
+
| kobest_hs_exact_match | 0.5500 |
|
52 |
+
| kobest_sn_exact_match | 0.9800 |
|
53 |
+
| kobest_wic_exact_match | 0.6200 |
|
54 |
+
| korea_cg_bleu | 0.0865 |
|
55 |
+
| kornli_exact_match | 0.6400 |
|
56 |
+
| korsts_pearson | 0.8547 |
|
57 |
+
| korsts_spearman| 0.8464 |
|
58 |
|
59 |
+
LogicKor
|
60 |
+
|
61 |
+
| ์นดํ
๊ณ ๋ฆฌ | ์ฑ๊ธ ์ ์ ํ๊ท | ๋ฉํฐ ์ ์ ํ๊ท |
|
62 |
+
|----------|------------------|-------------------|
|
63 |
+
| ์ฝ๋ฉ(Coding) | 7.71 | 7.71 |
|
64 |
+
| ์ํ(Math) | 5.57 | 3.86 |
|
65 |
+
| ์ดํด(Understanding) | 6.86 | 8.14 |
|
66 |
+
| ์ถ๋ก (Reasoning) | 8.14 | 6.43 |
|
67 |
+
| ๊ธ์ฐ๊ธฐ(Writing) | 8.71 | 6.86 |
|
68 |
+
| ๋ฌธ๋ฒ(Grammar) | 5.29 | 2.29 |
|
|
|
69 |
|
70 |
+
| ์นดํ
๊ณ ๋ฆฌ | ์ฑ๊ธ ์ ์ ํ๊ท | ๋ฉํฐ ์ ์ ํ๊ท |
|
71 |
+
|------------|------------------|-------------------|
|
72 |
+
| ์ ์ฒด ์ฑ๊ธ | 7.05 | 5.88 |
|
73 |
|
74 |
### Configuration
|
75 |
|