Edit model card

Built with Axolotl

kurugai/Kurugai-EEVE-v1.0๋Š” yanolja/EEVE-Korean-10.8B-v1.0๋ฅผ ๋ฒ ์ด์Šค๋ชจ๋ธ๋กœ ํ•ด์„œ maywell/kiqu_samples ๋ฐ์ดํ„ฐ์…‹์œผ๋กœ ํ•™์Šต๋œ ๋ชจ๋ธ์ž…๋‹ˆ๋‹ค.

ํ•™์Šต์‹œ๊ฐ„

RTX 8000 GPU 1EA๋กœ 16์‹œ๊ฐ„ ํ•™์Šตํ•˜์˜€์Šต๋‹ˆ๋‹ค.

๋ฒค์น˜๋งˆํฌ

Ko-LLM-Leaderboard

Average Ko-ARC Ko-HellaSwag Ko-MMLU Ko-TruthfulQA Ko-CommonGen V2
54.72 49.66 58.91 52.17 50.87 61.98

LogicKor ํ•œ๊ตญ์–ด ์–ธ์–ด๋ชจ๋ธ ๋‹ค๋ถ„์•ผ ์‚ฌ๊ณ ๋ ฅ ๋ฒค์น˜๋งˆํฌ

์ถ”๋ก  ์ˆ˜ํ•™ ๊ธ€์“ฐ๊ธฐ ์ฝ”๋”ฉ ์ดํ•ด ๋ฌธ๋ฒ• ์‹ฑ๊ธ€ํ„ด ๋ฉ€ํ‹ฐํ„ด ์ด์ 
3.00 2.71 3.92 4.21 6.35 4.14 4.59 3.52 4.05

ํƒฌํ”Œ๋ฆฟ

๋ณธ ๋ชจ๋ธ์„ ์•ŒํŒŒ์นด ํฌ๋งท์„ ๋”ฐ๋ฆ…๋‹ˆ๋‹ค.

Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.

### Instruction:
{instruction}

### Response:
{output}
Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.

### Instruction:
์ฝ”๋กœ๋‚˜19์— ๊ฑธ๋ ธ๋Š”๋ฐ ์–ด๋–ป๊ฒŒ ๋Œ€์ฒ˜ํ•ด์•ผํ•ด?

### Response:
 ์ฝ”๋กœ๋‚˜19์— ๊ฑธ๋ฆด ๊ฒฝ์šฐ ๊ฐ€์žฅ ์ค‘์š”ํ•œ ๊ฒƒ์€ ์ž์‹ ์˜ ๊ฑด๊ฐ• ์ƒํƒœ๋ฅผ ์ž˜ ๋ชจ๋‹ˆํ„ฐ๋งํ•˜๊ณ  ์ ์ ˆํ•œ ์˜๋ฃŒ ์กฐ์น˜๋ฅผ ์ทจํ•˜๋Š” ๊ฒƒ์ด์—์š”. ์—ฌ๊ธฐ ๋ช‡ ๊ฐ€์ง€ ๊ธฐ๋ณธ์ ์ธ ๋Œ€์‘ ๋ฐฉ๋ฒ•์„ ์†Œ๊ฐœํ• ๊ฒŒ์š”:

1. **์ฆ์ƒ ๊ด€๋ฆฌ**: ๋ฐœ์—ด, ๊ธฐ์นจ, ํ˜ธํก๊ณค๋ž€ ๋“ฑ์˜ ์ฆ์ƒ์ด ๋‚˜ํƒ€๋‚  ์ˆ˜ ์žˆ์–ด์š”. ์ด๋Ÿฌํ•œ ์ฆ์ƒ๋“ค์ด ์‹ฌํ•ด์ง€๊ฑฐ๋‚˜ ์•…ํ™”๋  ๊ฒฝ์šฐ ์ฆ‰์‹œ ์˜์‚ฌ์™€ ์ƒ๋‹ดํ•˜๊ฑฐ๋‚˜ ๋ณ‘์›์„ ๋ฐฉ๋ฌธํ•˜์„ธ์š”. ๋˜ํ•œ ์ถฉ๋ถ„ํ•œ ์ˆ˜๋ถ„ ์„ญ์ทจ์™€ ์˜์–‘ ๋ณด์ถฉ์„ ํ†ตํ•ด ๋ชธ์˜ ํšŒ๋ณต์„ ๋„์™€์ฃผ์„ธ์š”.

2. **์ ‘์ด‰์ž ์ถ”์  ๋ฐ ๊ฒฉ๋ฆฌ**: ๊ฐ์—ผ ์‚ฌ์‹ค์„ ์•Œ๊ฒŒ ๋˜๋ฉด ์ ‘์ด‰์ž๋“ค๊ณผ์˜ ์†Œํ†ต์„ ์‹œ์ž‘ํ•˜๊ณ  ํ•„์š”ํ•œ ๊ฒฝ์šฐ ์ž๊ฐ€๊ฒฉ๋ฆฌ๋ฅผ ์‹œ์ž‘ํ•˜์…”์•ผ ํ•ด์š”. ์ด๋Š” ๋‹ค๋ฅธ ์‚ฌ๋žŒ๋“ค์—๊ฒŒ ๋ฐ”์ด๋Ÿฌ์Šค๋ฅผ ์ „ํŒŒํ•˜์ง€ ์•Š๋„๋ก ์˜ˆ๋ฐฉํ•˜๊ธฐ ์œ„ํ•œ ์กฐ์น˜์˜ˆ์š”.

3. **์˜๋ฃŒ ์„œ๋น„์Šค ํ™œ์šฉ**: ์ฆ์ƒ์ด ์‹ฌ๊ฐํ•ด์งˆ ๊ฒฝ์šฐ ์ฆ‰๊ฐ์ ์ธ ์˜๋ฃŒ ๋„์›€์„ ๋ฐ›์•„์•ผ ํ•ด์š”. ๋ณ‘์›์— ๊ฐ€์„œ ์ง„๋ฃŒ๋ฅผ ๋ฐ›๊ฑฐ๋‚˜, ํ•„์š”ํ•˜๋‹ค๋ฉด ์ž…์›์„ ๊ณ ๋ คํ•ด๋ณด์„ธ์š”. ๋˜ํ•œ ์ฒ˜๋ฐฉ๋œ ์•ฝ๋ฌผ์ด๋‚˜ ์น˜๋ฃŒ ๊ณ„ํš์„ ๋”ฐ๋ฅด๋Š” ๊ฒƒ๋„ ์ค‘์š”ํ•ด์š”.

4. **์ •์‹ ์  ์ง€์› ๋ฐ›๊ธฐ**: ๊ฐ์—ผ๋ณ‘์œผ๋กœ ์ธํ•œ ์ŠคํŠธ๋ ˆ์Šค๋‚˜ ๋ถˆ์•ˆ์€ ์ž์—ฐ์Šค๋Ÿฌ์šด ๋ฐ˜์‘์ด์—์š”. ๋งŒ์•ฝ ํ•„์š”ํ•˜๋ฉด ์ •์‹ ๊ฑด๊ฐ• ์ „๋ฌธ๊ฐ€์˜ ์ƒ๋‹ด์„ ๋ฐ›์•„ ๋ณด์‹œ๋Š” ๊ฒƒ์„ ๊ถŒ์žฅ๋“œ๋ ค์š”. ์˜จ๋ผ์ธ ์ƒ๋‹ด๋„ ๊ฐ€๋Šฅํ•˜๋‹ˆ ์ฐธ๊ณ ํ•ด ์ฃผ์„ธ์š”.

5. **์˜ˆ๋ฐฉ ์กฐ์น˜ ์œ ์ง€ํ•˜๊ธฐ**: ๋ฐฑ์‹  ์ ‘์ข…, ๋งˆ์Šคํฌ ์ฐฉ์šฉ, ์† ์”ป๊ธฐ ๋“ฑ ์ฝ”๋กœ๋‚˜19 ์˜ˆ๋ฐฉ์„ ์œ„ํ•œ ๊ธฐ๋ณธ ๊ทœ์น™์„ ๊ณ„์† ์œ ์ง€ํ•ด์ฃผ์„ธ์š”. ์ด๋ฅผ ํ†ตํ•ด ์ž์‹ ๊ณผ ์ฃผ๋ณ€ ์‚ฌ๋žŒ๋“ค์˜ ๊ฑด๊ฐ•์„ ๋ณดํ˜ธํ•  ์ˆ˜ ์žˆ์–ด์š”.

6. **์ง€์†์ ์ธ ์ •๋ณด ์—…๋ฐ์ดํŠธ**: ์งˆ๋ณ‘๊ด€๋ฆฌ์ฒญ์ด๋‚˜ ๊ด€๋ จ ์ •๋ถ€ ๊ธฐ๊ด€์—์„œ ์ œ๊ณตํ•˜๋Š” ์ตœ์‹  ์ •๋ณด๋ฅผ ์ˆ˜์‹œ๋กœ ํ™•์ธํ•ด ๋ณด์„ธ์š”. ์ด๋ ‡๊ฒŒ ํ•˜๋ฉด ์ž์‹ ์—๊ฒŒ ๋งž๋Š” ๋Œ€์‘์„ ํ•  ์ˆ˜ ์žˆ๋Š” ์ •๋ณด๋ฅผ ์–ป์„ ์ˆ˜ ์žˆ์–ด์š”.

์ฝ”๋กœ๋‚˜19๋Š” ๊ฐœ์ธ๋งˆ๋‹ค ๋‹ค๋ฅด๊ฒŒ ์˜ํ–ฅ์„ ๋ฏธ์น  ์ˆ˜ ์žˆ์œผ๋‹ˆ, ๊ฐœ์ธ์˜ ์ƒํ™ฉ์— ๋งž๊ฒŒ ์ ์ ˆํžˆ ๋Œ€์‘ํ•  ํ•„์š”๊ฐ€ ์žˆ์–ด์š”. ํ•ญ์ƒ ์ „๋ฌธ๊ฐ€์™€์˜ ์ƒ์˜๋ฅผ ํ†ตํ•œ ์ ‘๊ทผ์ด ์ตœ์„ ์ด๋ผ๊ณ  ํ•  ์ˆ˜ ์žˆ๊ฒ ์ฃ . ๋” ๊ถ๊ธˆํ•œ ์ ์ด ์žˆ์œผ๋ฉด ์–ธ์ œ๋“ ์ง€ ๋ฌผ์–ด๋ด ์ฃผ์„ธ์š”!

๋„์›€์„ ์ฃผ์‹ ๋ถ„

์ด ๋ชจ๋ธ์€ ์•„๋‚ด์˜ ์ง€์›์„ ๋ฐ›์•„ ์ œ์ž‘๋˜์—ˆ์Šต๋‹ˆ๋‹ค. ์•„๋‚ด์—๊ฒŒ ๊ฐ์‚ฌ์˜ ๋ง์„ ์ „ํ•ฉ๋‹ˆ๋‹ค.

Downloads last month
3,222
Safetensors
Model size
10.8B params
Tensor type
FP16
ยท
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for kurugai/Kurugai-EEVE-v1.0

Finetuned
(9)
this model
Finetunes
1 model

Dataset used to train kurugai/Kurugai-EEVE-v1.0