Llama-3-Instruct-8B-DPO-v0.2: benchmark results
Princeton NLP's DPO v0.2 baseline on Llama 3 8B Instruct, released with the SimPO paper's updated baseline suite. Provider: Princeton NLP. Released 2024-07-06. Access: Open.
Aggregate Rating 1442 ± 19, rank #1042 of 1661 rated models, from 21 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-GPQA-Diamond | 27.78 | Normalized accuracy (%) | 95 |
| Open Korean LLM - KorNAT-Helpful | 51.45 | Normalized accuracy (%) | 90.7 |
| Open LLM Leaderboard - IFEval | 72.08 | Score | 85.9 |
| Open Korean LLM - Ko-IFEval | 40.14 | Strict accuracy, prompt/instruction mean (%) | 82.1 |
| Open Korean LLM - Ko-EQ-Bench | 31.52 | EQ-Bench score | 80.8 |
| Open Korean LLM Leaderboard | 41.13 | Average Score (%) | 80.8 |
| Open Ko-LLM Leaderboard | 41.13 | Average (%) | 80.7 |
| Open Portuguese LLM - OAB Exams | 50.52 | Accuracy (%) | 77.3 |
| Open Portuguese LLM - BLUEX | 58.28 | Accuracy (%) | 72.6 |
| Open Portuguese LLM - ENEM | 69 | Accuracy (%) | 71.7 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 36.09 | Exact match, flexible extract (%) | 69.4 |
| Open Korean LLM - KorNAT-CKA | 26.48 | Normalized accuracy (%) | 64.9 |
Interactive version: theaggregate.ai/model?slug=llama-3-instruct-8b-dpo-v0-2 · How It Works · Data refreshed daily, snapshot 2026-10-23.