Barcenas-Llama3-8B-ORPO: benchmark results
Danielbrdz's ORPO fine-tune of VAGO's Llama-3-SauerkrautLM-8B-Instruct, trained on the dolphin-sft-v0.1-preference dataset. Provider: Danielbrdz. Released 2024-05-01. Access: Open.
Aggregate Rating 1458 ± 14, rank #949 of 1656 rated models, from 46 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Medical LLM - PubMedQA | 77.6 | Accuracy (%) | 94 |
| Open Korean LLM - Ko-EQ-Bench | 40.49 | EQ-Bench score | 91.8 |
| Open Korean LLM - KorNAT-Helpful | 51.4 | Normalized accuracy (%) | 90.4 |
| Open LLM Leaderboard v1 - MMLU | 68.02 | Accuracy (%) (5-shot) | 90.3 |
| Open Medical LLM - MMLU College Biology | 81.25 | Accuracy (%) | 89.6 |
| Open LLM Leaderboard v1 - ARC Challenge | 71.42 | Normalized accuracy (%) (25-shot) | 89.2 |
| Open LLM Leaderboard - IFEval | 73.72 | Score | 88.3 |
| Open Korean LLM Leaderboard | 42.34 | Average Score (%) | 86 |
| Open Ko-LLM Leaderboard | 42.34 | Average (%) | 85.9 |
| Open Medical LLM - MMLU Medical Genetics | 83 | Accuracy (%) | 84.1 |
| Open LLM Leaderboard v1 - HellaSwag | 86.84 | Normalized accuracy (%) (10-shot) | 83.2 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 48.29 | Exact match, flexible extract (%) | 81.9 |
Interactive version: theaggregate.ai/model?slug=barcenas-llama3-8b-orpo · How It Works · Data refreshed daily, snapshot 2026-10-20.