Aura-llama-3
Now that the cute anime girl has your attention.
UPDATE: Model has been fixed
Aura-llama is using the methodology presented by SOLAR for scaling LLMs called depth up-scaling (DUS), which encompasses architectural modifications with continued pretraining. Using the solar paper as a base, I integrated Llama-3 weights into the upscaled layers, and In the future plan to continue training the model.
Aura-llama is a merge of the following models to create a base model to work from:
Merged Evals (Has Not Been Finetuned):
Aura-llama
- Avg: 63.13
- ARC: 58.02
- HellaSwag: 77.82
- MMLU: 65.61
- T-QA: 51.94
- Winogrande: 73.40
- GSM8K: 52.01
🧩 Configuration
dtype: float16
merge_method: passthrough
slices:
- sources:
- layer_range: [0, 12]
model: NousResearch/Meta-Llama-3-8B-Instruct
- sources:
- layer_range: [8, 20]
model: NousResearch/Meta-Llama-3-8B-Instruct
- sources:
- layer_range: [16, 28]
model: NousResearch/Meta-Llama-3-8B-Instruct
- sources:
- layer_range: [24, 32]
model: NousResearch/Meta-Llama-3-8B-Instruct
Open LLM Leaderboard Evaluation Results
Detailed results can be found here
| Metric | Value |
|---|---|
| Avg. | 63.13 |
| AI2 Reasoning Challenge (25-Shot) | 58.02 |
| HellaSwag (10-Shot) | 77.82 |
| MMLU (5-Shot) | 65.61 |
| TruthfulQA (0-shot) | 51.94 |
| Winogrande (5-shot) | 73.40 |
| GSM8k (5-shot) | 52.01 |
- Downloads last month
- 937
Finetuned from
Evaluation results
- normalized accuracy on AI2 Reasoning Challenge (25-Shot)test set Open LLM Leaderboard58.020
- normalized accuracy on HellaSwag (10-Shot)validation set Open LLM Leaderboard77.820
- accuracy on MMLU (5-Shot)test set Open LLM Leaderboard65.610
- mc2 on TruthfulQA (0-shot)validation set Open LLM Leaderboard51.940
- accuracy on Winogrande (5-shot)validation set Open LLM Leaderboard73.400
- accuracy on GSM8k (5-shot)test set Open LLM Leaderboard52.010