mistral-orpo-capybara-7k Benchmark Results

mistral-orpo-capybara-7k — benchmark results

Provider: Mistral. Released 2024-03-23. Access: Open.

Unified ELO 1351 ± 18, rank #1409 of 1776 rated models, from 15 benchmark results.

Benchmark	Score	Metric	Percentile
Open LLM Leaderboard - IFEval	53.67	Score	64.2
Open LLM Leaderboard - GPQA	4.81	Score	41
Open Korean LLM Leaderboard	34.35	Average Score (%)	35.7
Open LLM Leaderboard - MMLU-Pro	21.9	Score	35.5
Open LLM Leaderboard - MuSR	7.58	Score	35.4
Open LLM Leaderboard - BBH	23.43	Score	35.3
Open Chinese LLM - TruthfulQA MC	51.8	Accuracy (%)	32.3
Open Chinese LLM - ARC Challenge	47.78	Accuracy (%)	31.6
Open Chinese LLM - HellaSwag	55.9	Accuracy (%)	28
Open Chinese LLM - CMMLU	47.89	Accuracy (%)	27.9
Open LLM Leaderboard - MATH Level 5	3.93	Score	22.6
Open Chinese LLM - GSM8K	27.9	Accuracy (%)	20.8