ELYZA-japanese-Llama-2-13B-fast-instruct: benchmark results
Provider: Other. Access: Open.
Unified ELO 1395 ± 21, rank #2181 of 2928 rated models, from 14 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - HellaSwag | 81.82 | Normalized accuracy (%) (10-shot) | 47.2 |
| Open LLM Leaderboard v1 - WinoGrande | 75.93 | Accuracy (%) (5-shot) | 45 |
| Open LLM Leaderboard v1 - ARC Challenge | 57.51 | Normalized accuracy (%) (25-shot) | 37.7 |
| pfgen-bench - Completion Mode - Truthfulness | 0.64 | Truthfulness Score | 36.4 |
| pfgen-bench - Completion Mode - Helpfulness | 0.07 | Helpfulness Score | 34.7 |
| pfgen-bench - Completion Mode - Fluency | 0.5 | Fluency Score | 34.5 |
| Open LLM Leaderboard v1 - MMLU | 54.52 | Accuracy (%) (5-shot) | 34.1 |
| pfgen-bench - Completion Mode - Score | 0.4 | pfgen Score (mean of three) | 33.9 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 43.82 | MC2 (%) (0-shot) | 24.9 |
| Open LLM Leaderboard v1 - GSM8K | 2.73 | Accuracy (%) (5-shot) | 20.7 |
| pfgen-bench - QA Mode - Helpfulness | 0.05 | Helpfulness Score | 17.9 |
| pfgen-bench - QA Mode - Score | 0.33 | pfgen Score (mean of three) | 14.5 |
Interactive version: theaggregate.ai/model?slug=elyza-japanese-llama-2-13b-fast-instruct · How It Works · Data refreshed daily, snapshot 2026-09-23.