gemma-2-27B-it-abliterated: benchmark results
Provider: Google. Released 2024-08-28. Access: API.
Unified ELO 1659 ± 17, rank #428 of 2656 rated models, from 15 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Ko-LLM Leaderboard | 55.5 | Average (%) | 99.9 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 78.01 | Exact match, flexible extract (%) | 99.9 |
| Open Korean LLM - Ko-IFEval | 72.87 | Strict accuracy, prompt/instruction mean (%) | 99.9 |
| Open Korean LLM Leaderboard | 55.5 | Average Score (%) | 99.9 |
| Open Korean LLM - KorNAT-CKA | 38.75 | Normalized accuracy (%) | 98.1 |
| Open Korean LLM - Ko-EQ-Bench | 57.84 | EQ-Bench score | 97.4 |
| Open Korean LLM - Ko-Winogrande | 69.69 | Accuracy (%) | 95.4 |
| Open Portuguese LLM - BLUEX | 71.07 | Accuracy (%) | 92.5 |
| Open Korean LLM - Ko-GPQA-Diamond | 27.27 | Normalized accuracy (%) | 92.4 |
| Open Portuguese LLM - ENEM | 78.59 | Accuracy (%) | 92.1 |
| Open Portuguese LLM - OAB Exams | 57.77 | Accuracy (%) | 90.9 |
| Open Korean LLM - KorNAT-Helpful | 50.52 | Normalized accuracy (%) | 86.2 |
Interactive version: theaggregate.ai/model?slug=gemma-2-27b-it-abliterated · How It Works · Data refreshed daily, snapshot 2026-09-19.