Spaces:
Sleeping
Sleeping
| HEAD_TEXT = """ | |
| This is the official leaderboard for 🏅StructEval benchmark. Starting from an atomic test objective, StructEval deepens and broadens the evaluation by conducting a **structured assessment across multiple cognitive levels and critical concepts**, and therefore offers a comprehensive, robust and consistent evaluation for LLMs. | |
| Please refer to 🐱[StructEval repository](https://github.com/c-box/StructEval) for model evaluation and 📖[our paper]() for experimental analysis. | |
| 🚀 **_Latest News_** | |
| * [2024.8.2] We released the first version of StructEval leaderboard, which includes 21 open-sourced language models, more datasets and models as comming soon🔥🔥🔥. | |
| * [2024.7.31] We regenerated the StructEval Benchmark based on the latest [Wikipedia](https://www.wikipedia.org/) pages (20240601) using [GPT-4o-mini](https://openai.com/index/gpt-4o-mini-advancing-cost-efficient-intelligence/) model, which could minimize the impact of data contamination🔥🔥🔥. | |
| """ | |
| ABOUT_TEXT = """# What is StructEval? | |
| Evaluation is the baton for the development of large language models. | |
| Current evaluations typically employ *a single-item assessment paradigm* for each atomic test objective, which struggles to discern whether a model genuinely possesses the required capabilities or merely memorizes/guesses the answers to specific questions. | |
| To this end, we propose a novel evaluation framework referred to as ***StructEval***. | |
| Starting from an atomic test objective, StructEval deepens and broadens the evaluation by conducting a **structured assessment across multiple cognitive levels and critical concepts**, and therefore offers a comprehensive, robust and consistent evaluation for LLMs. | |
| Experiments demonstrate that **StructEval serves as a reliable tool for resisting the risk of data contamination and reducing the interference of potential biases**, thereby providing more reliable and consistent conclusions regarding model capabilities. | |
| Our framework also sheds light on the design of future principled and trustworthy LLM evaluation protocols. | |
| # How to evaluate? | |
| Our 🐱[repo](https://github.com/c-box/StructEval) provides easy-to-use scripts for both evaluating LLMs on existing StructEval benchmarks and generating new benchmarks based on StructEval framework. | |
| # Contact | |
| If you have any questions, feel free to reach out to us at [boxi2020@iscas.ac.cn](mailto:boxi2020@iscas.ac.cn). | |
| """ | |
| CITATION_BUTTON_LABEL = "Copy the following snippet to cite these results" | |
| CITATION_BUTTON_TEXT = r""" | |
| comming soon | |
| """ | |
| ACKNOWLEDGEMENT_TEXT = """ | |
| Inspired from the [🤗 Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard). | |
| """ | |
| NOTES_TEXT = """ | |
| * On most models on base MMLU, we collected the results for their official technical report. For the models that have not been reported, we use opencompass for evaluation. | |
| * For other 2 base benchmarks and all 3 structured benchmarks: for chat models, we evaluate them under 0-shot setting; for completion model, we evaluate them under 0-shot setting with ppl. | |
| """ |