openbuddy-llama3.1-8B-v22.3-131k Benchmark Results

openbuddy-llama3.1-8B-v22.3-131k: benchmark results

OpenBuddy's multilingual chatbot fine-tune of Llama 3.1 8B Instruct with a 131k context. Provider: OpenBuddy. Released 2024-08-16. Access: Open.

Aggregate Rating 1439 ± 19, rank #1058 of 1661 rated models, from 21 benchmark results.

Benchmark	Score	Metric	Percentile
Open Korean LLM - Ko-IFEval	43.77	Strict accuracy, prompt/instruction mean (%)	88.2
Open Korean LLM - Ko-EQ-Bench	34.07	EQ-Bench score	83.7
Open Korean LLM - Ko-GSM8K (flexible extract)	48.52	Exact match, flexible extract (%)	82.2
Open Korean LLM Leaderboard	40.38	Average Score (%)	76.7
Open Ko-LLM Leaderboard	40.38	Average (%)	76.6
Open Korean LLM - KorNAT-Harmless	64.73	Normalized accuracy (%)	76.6
Open Korean LLM - Ko-Winogrande	63.54	Accuracy (%)	74.7
Open LLM Leaderboard - IFEval	59.97	Score	71.8
Open Portuguese LLM - OAB Exams	48.25	Accuracy (%)	68.9
Open Portuguese LLM - ENEM	67.6	Accuracy (%)	67.4
Open Portuguese LLM - FaQuAD NLI	71.51	Macro F1 (%)	63.7
Open Portuguese LLM - BLUEX	55.49	Accuracy (%)	63.4