Llama 3 8B Instruct Code [4bit] [Adapter]: benchmark results
An ibivibiv LoRA adapter for meta-llama/Meta-Llama-3-8B-Instruct. Its embedded training recipe uses QLoRA with the base loaded in 4-bit and the candenizkocak/code-alpaca-297k dataset. The repository publishes adapter weights; this does not establish 4-bit storage of the adapter itself. Provider: ibivibiv. Access: Open.
Aggregate Rating 1394 ± 36, rank #6369 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - GSM8K | 58.07 | Accuracy (%) (5-shot) | 69.7 |
| Open LLM Leaderboard v1 - MMLU | 64.38 | Accuracy (%) (5-shot) | 69.2 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 48.84 | MC2 (%) (0-shot) | 42.8 |
| Open LLM Leaderboard v1 - ARC Challenge | 56.4 | Normalized accuracy (%) (25-shot) | 35.4 |
| Open LLM Leaderboard v1 - WinoGrande | 73.8 | Accuracy (%) (5-shot) | 33.6 |
| Open LLM Leaderboard v1 - HellaSwag | 78.32 | Normalized accuracy (%) (10-shot) | 33.5 |
Interactive version: theaggregate.ai/model?slug=llama-3-8b-instruct-code-4bit-adapter · How It Works · Data refreshed daily, snapshot 2026-11-08.