llama3_instruct_tuning_without_pretraing Benchmark Results

llama3_instruct_tuning_without_pretraing: benchmark results

Provider: Meta. Access: Open.

Unified ELO 1430 ± 17, rank #1666 of 2656 rated models, from 10 benchmark results.

Benchmark	Score	Metric	Percentile
Open Korean LLM - KorNAT-CKA	31.51	Normalized accuracy (%)	86.9
Open Korean LLM - Ko-GPQA-Diamond	25.76	Normalized accuracy (%)	81.3
Open Korean LLM - Ko-Winogrande	64.4	Accuracy (%)	78.4
Open Korean LLM - Ko-GSM8K (flexible extract)	23.65	Exact match, flexible extract (%)	56.7
Open Korean LLM - KorNAT-Harmless	63.05	Normalized accuracy (%)	54.5
Open Korean LLM - Ko-IFEval	29.84	Strict accuracy, prompt/instruction mean (%)	53.2
Open Korean LLM Leaderboard	35.42	Average Score (%)	51.5
Open Ko-LLM Leaderboard	35.42	Average (%)	51.4
Open Korean LLM - KorNAT-Helpful	46.65	Normalized accuracy (%)	48.1
Open Korean LLM - Ko-EQ-Bench	0	EQ-Bench score	24.5