---
license: apache-2.0
datasets:
- itsliupeng/mmlu_recall
language:
- en
pipeline_tag: text-generation
---

We are utilizing the [mmlu_recall](https://huggingface.co/datasets/itsliupeng/mmlu_recall) dataset to continuously train the [Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) model, aiming to enhance performance on mmlu metrics, while ensuring that other metric performances remain unaffected.

# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_itsliupeng__llama2_7b_mmlu)

| Metric                | Value                     |
|-----------------------|---------------------------|
| Avg.                  | 46.31   |
| ARC (25-shot)         | 56.14          |
| HellaSwag (10-shot)   | 79.13    |
| MMLU (5-shot)         | 60.04         |
| TruthfulQA (0-shot)   | 40.95   |
| Winogrande (5-shot)   | 74.43   |
| GSM8K (5-shot)        | 7.88        |
| DROP (3-shot)         | 5.59         |