Llama-3-Instruct-8B-CPO Benchmark Results

Llama-3-Instruct-8B-CPO — benchmark results

Provider: Meta. Released 2024-07-06. Access: Open.

Unified ELO 1350 ± 89, rank #1418 of 1776 rated models, from 7 benchmark results.

Benchmark	Score	Metric	Percentile
Open Korean LLM Leaderboard	548.28	Average Score (%)	91.8
Open LLM Leaderboard - IFEval	72.93	Score	87
Open LLM Leaderboard - MMLU-Pro	29.47	Score	57.1
Open LLM Leaderboard - BBH	28.6	Score	47.1
Open LLM Leaderboard - MATH Level 5	9.89	Score	47
Open LLM Leaderboard - GPQA	1.34	Score	13.2
Open LLM Leaderboard - MuSR	1.76	Score	7.2