k2s3_test_24001 [float16] [rev: 60f591877327]: benchmark results
Changgil Song release published through Changgil: a PEFT LoRA fine-tune of meta-llama/Llama-2-13b-chat-hf. This catalog row remains bound to revision 60f5918773275ff16e43a945a24dd4ad8ddfeacf and its explicit float16 evaluation mode. LoRA training does not imply that the distributed files are adapter-only. Provider: Changgil. Access: Open.
Aggregate Rating 1350 ± 33, rank #7686 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - GSM8K | 29.87 | Accuracy (%) (5-shot) | 47 |
| Open LLM Leaderboard v1 - WinoGrande | 75.69 | Accuracy (%) (5-shot) | 43.5 |
| Open LLM Leaderboard v1 - HellaSwag | 80.59 | Normalized accuracy (%) (10-shot) | 41.5 |
| Open LLM Leaderboard v1 - ARC Challenge | 55.8 | Normalized accuracy (%) (25-shot) | 34.2 |
| Open LLM Leaderboard v1 - MMLU | 54.42 | Accuracy (%) (5-shot) | 33.9 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 43.62 | MC2 (%) (0-shot) | 24.2 |
Interactive version: theaggregate.ai/model?slug=k2s3-test-24001-float16-rev-60f591877327 · How It Works · Data refreshed daily, snapshot 2026-11-08.