Fireball-Alpaca-Llama3.1.07-8B-Philos-Math-KTO-beta: benchmark results
EpistemeAI's KTO-aligned Llama 3.1 8B Instruct fine-tune from its Fireball Alpaca line, targeting philosophy and math reasoning. Provider: EpistemeAI. Released 2024-09-12. Access: Open.
Aggregate Rating 1435 ± 21, rank #1071 of 1654 rated models, from 20 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-IFEval | 50.47 | Strict accuracy, prompt/instruction mean (%) | 93.9 |
| Open Korean LLM - Ko-EQ-Bench | 36.51 | EQ-Bench score | 86.9 |
| Open LLM Leaderboard - IFEval | 72.74 | Score | 86.6 |
| Open Portuguese LLM - ASSIN2 RTE | 92.77 | Macro F1 (%) | 83.4 |
| Open Korean LLM Leaderboard | 40.06 | Average Score (%) | 75.3 |
| Open Korean LLM - KorNAT-Helpful | 48.78 | Normalized accuracy (%) | 72.3 |
| Open Portuguese LLM - OAB Exams | 49.34 | Accuracy (%) | 72.1 |
| Open Korean LLM - Ko-GPQA-Diamond | 24.75 | Normalized accuracy (%) | 69.5 |
| Open Portuguese LLM - ENEM | 68.23 | Accuracy (%) | 69.3 |
| Open Portuguese LLM - BLUEX | 56.88 | Accuracy (%) | 67.6 |
| Open Korean LLM - KorNAT-Harmless | 63.64 | Normalized accuracy (%) | 65.1 |
| Open Portuguese LLM - FaQuAD NLI | 71.65 | Macro F1 (%) | 64.2 |
Interactive version: theaggregate.ai/model?slug=fireball-alpaca-llama3-1-07-8b-philos-math-kto-beta · How It Works · Data refreshed daily, snapshot 2026-10-24.