Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge: benchmark results
Community SLERP merge by grimjim of UCLA-AGI's SPPO Iter3 and Princeton NLP's SimPO fine-tunes of Llama 3 8B Instruct. Provider: grimjim. Released 2024-06-28. Access: Open.
Aggregate Rating 1433 ± 17, rank #1086 of 1656 rated models, from 29 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - KorNAT-Helpful | 50.83 | Normalized accuracy (%) | 87.3 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 52.69 | Exact match, flexible extract (%) | 86.1 |
| Open Korean LLM - Ko-IFEval | 42.24 | Strict accuracy, prompt/instruction mean (%) | 86 |
| Open Korean LLM - Ko-EQ-Bench | 29.64 | EQ-Bench score | 78.1 |
| Open Korean LLM - Ko-GPQA-Diamond | 25.25 | Normalized accuracy (%) | 75.9 |
| Open Korean LLM - KorNAT-Harmless | 64.48 | Normalized accuracy (%) | 74.5 |
| Open Korean LLM Leaderboard | 39.7 | Average Score (%) | 73.6 |
| Open Ko-LLM Leaderboard | 39.7 | Average (%) | 73.4 |
| Open Portuguese LLM - OAB Exams | 49.48 | Accuracy (%) | 72.9 |
| Open Portuguese LLM - ENEM | 69.14 | Accuracy (%) | 72.4 |
| Open Chinese LLM - TruthfulQA MC | 56.61 | Accuracy (%) | 70.7 |
| Open Portuguese LLM - BLUEX | 57.3 | Accuracy (%) | 69.4 |
Interactive version: theaggregate.ai/model?slug=llama-3-instruct-8b-sppo-iter3-simpo-merge · How It Works · Data refreshed daily, snapshot 2026-10-20.