Llama-3-constructionsafety-LoRA: benchmark results
Provider: Meta. Access: Open.
Unified ELO 1432 ± 18, rank #1650 of 2656 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - KorNAT-CKA | 34.82 | Normalized accuracy (%) | 94.9 |
| Open Korean LLM - Ko-Winogrande | 65.35 | Accuracy (%) | 84 |
| Open Korean LLM - Ko-GPQA-Diamond | 25.25 | Normalized accuracy (%) | 75.9 |
| Open Korean LLM - KorNAT-Helpful | 47.93 | Normalized accuracy (%) | 62.5 |
| Open Ko-LLM Leaderboard | 35.98 | Average (%) | 55.2 |
| Open Korean LLM Leaderboard | 35.97 | Average Score (%) | 55.1 |
| Open Korean LLM - Ko-EQ-Bench | 7.61 | EQ-Bench score | 53.9 |
| Open Korean LLM - Ko-IFEval | 29.29 | Strict accuracy, prompt/instruction mean (%) | 51 |
| Open Korean LLM - KorNAT-Harmless | 62.39 | Normalized accuracy (%) | 43.3 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 0.08 | Exact match, flexible extract (%) | 4.8 |
Interactive version: theaggregate.ai/model?slug=llama-3-constructionsafety-lora · How It Works · Data refreshed daily, snapshot 2026-09-19.