mirror of
https://github.com/open-compass/opencompass.git
synced 2025-05-30 16:03:24 +08:00
12b84aeb3b
* update gemini api and add gemini models * add openai models * update CHARM evaluation * add CHARM memorization tasks * add CharmMemSummarizer (output eval details for memorization-independent reasoning analysis * update CHARM readme --------- Co-authored-by: wujiang <wujiang@pjlab.org.cn>
165 lines
8.7 KiB
Markdown
165 lines
8.7 KiB
Markdown
# CHARM✨ Benchmarking Chinese Commonsense Reasoning of LLMs: From Chinese-Specifics to Reasoning-Memorization Correlations [ACL2024]
|
|
[](https://arxiv.org/abs/2403.14112)
|
|
[](./LICENSE)
|
|
<div align="center">
|
|
|
|
📃[Paper](https://arxiv.org/abs/2403.14112)
|
|
🏰[Project Page](https://opendatalab.github.io/CHARM/)
|
|
🏆[Leaderboard](https://opendatalab.github.io/CHARM/leaderboard.html)
|
|
✨[Findings](https://opendatalab.github.io/CHARM/findings.html)
|
|
|
|
</div>
|
|
|
|
<div align="center">
|
|
📖 <a href="./README_ZH.md"> 中文</a> | <a href="./README.md">English</a>
|
|
</div>
|
|
|
|
## Dataset Description
|
|
|
|
**CHARM** is the first benchmark for comprehensively and in-depth evaluating the commonsense reasoning ability of large language models (LLMs) in Chinese, which covers both globally known and Chinese-specific commonsense. In addition, the CHARM can evaluate the LLMs' memorization-independent reasoning abilities and analyze the typical errors.
|
|
|
|
## Comparison of commonsense reasoning benchmarks
|
|
<html lang="en">
|
|
<table align="center">
|
|
<thead class="fixed-header">
|
|
<tr>
|
|
<th>Benchmarks</th>
|
|
<th>CN-Lang</th>
|
|
<th>CSR</th>
|
|
<th>CN-specifics</th>
|
|
<th>Dual-Domain</th>
|
|
<th>Rea-Mem</th>
|
|
</tr>
|
|
</thead>
|
|
<tr>
|
|
<td>Most benchmarks in <a href="https://arxiv.org/abs/2302.04752"> davis2023benchmarks</a></td>
|
|
<td><strong><span style="color: red;">✘</span></strong></td>
|
|
<td><strong><span style="color: green;">✔</span></strong></td>
|
|
<td><strong><span style="color: red;">✘</span></strong></td>
|
|
<td><strong><span style="color: red;">✘</span></strong></td>
|
|
<td><strong><span style="color: red;">✘</span></strong></td>
|
|
</tr>
|
|
<tr>
|
|
<td><a href="https://arxiv.org/abs/1809.05053"> XNLI</a>, <a
|
|
href="https://arxiv.org/abs/2005.00333">XCOPA</a>,<a
|
|
href="https://arxiv.org/abs/2112.10668">XStoryCloze</a></td>
|
|
<td><strong><span style="color: green;">✔</span></strong></td>
|
|
<td><strong><span style="color: green;">✔</span></strong></td>
|
|
<td><strong><span style="color: red;">✘</span></strong></td>
|
|
<td><strong><span style="color: red;">✘</span></strong></td>
|
|
<td><strong><span style="color: red;">✘</span></strong></td>
|
|
</tr>
|
|
<tr>
|
|
<td><a href="https://arxiv.org/abs/2007.08124">LogiQA</a>, <a
|
|
href="https://arxiv.org/abs/2004.05986">CLUE</a>, <a
|
|
href="https://arxiv.org/abs/2306.09212">CMMLU</a></td>
|
|
<td><strong><span style="color: green;">✔</span></strong></td>
|
|
<td><strong><span style="color: red;">✘</span></strong></td>
|
|
<td><strong><span style="color: green;">✔</span></strong></td>
|
|
<td><strong><span style="color: red;">✘</span></strong></td>
|
|
<td><strong><span style="color: red;">✘</span></strong></td>
|
|
</tr>
|
|
<tr>
|
|
<td><a href="https://arxiv.org/abs/2312.12853">CORECODE</a> </td>
|
|
<td><strong><span style="color: green;">✔</span></strong></td>
|
|
<td><strong><span style="color: green;">✔</span></strong></td>
|
|
<td><strong><span style="color: red;">✘</span></strong></td>
|
|
<td><strong><span style="color: red;">✘</span></strong></td>
|
|
<td><strong><span style="color: red;">✘</span></strong></td>
|
|
</tr>
|
|
<tr>
|
|
<td><strong><a href="https://arxiv.org/abs/2403.14112">CHARM (ours)</a> </strong></td>
|
|
<td><strong><span style="color: green;">✔</span></strong></td>
|
|
<td><strong><span style="color: green;">✔</span></strong></td>
|
|
<td><strong><span style="color: green;">✔</span></strong></td>
|
|
<td><strong><span style="color: green;">✔</span></strong></td>
|
|
<td><strong><span style="color: green;">✔</span></strong></td>
|
|
</tr>
|
|
</table>
|
|
|
|
"CN-Lang" indicates the benchmark is presented in Chinese language. "CSR" means the benchmark is designed to focus on <strong>C</strong>ommon<strong>S</strong>ense <strong>R</strong>easoning. "CN-specific" indicates the benchmark includes elements that are unique to Chinese culture, language, regional characteristics, history, etc. "Dual-Domain" indicates the benchmark encompasses both Chinese-specific and global domain tasks, with questions presented in the similar style and format. "Rea-Mem" indicates the benchmark includes closely-interconnected <strong>rea</strong>soning and <strong>mem</strong>orization tasks.
|
|
|
|
|
|
## 🛠️ How to Use
|
|
Below are the steps for quickly downloading CHARM and using OpenCompass for evaluation.
|
|
|
|
### 1. Download CHARM
|
|
```bash
|
|
git clone https://github.com/opendatalab/CHARM ${path_to_CHARM_repo}
|
|
|
|
cd ${path_to_opencompass}
|
|
mkdir data
|
|
ln -snf ${path_to_CHARM_repo}/data/CHARM ./data/CHARM
|
|
```
|
|
### 2. Run Inference and Evaluation
|
|
```bash
|
|
cd ${path_to_opencompass}
|
|
|
|
# modify config file `configs/eval_charm_rea.py`: uncomment or add models you want to evaluate
|
|
python run.py configs/eval_charm_rea.py -r --dump-eval-details
|
|
|
|
# modify config file `configs/eval_charm_mem.py`: uncomment or add models you want to evaluate
|
|
python run.py configs/eval_charm_mem.py -r --dump-eval-details
|
|
```
|
|
The inference and evaluation results would be in `${path_to_opencompass}/outputs`, like this:
|
|
```bash
|
|
outputs
|
|
├── CHARM_mem
|
|
│ └── chat
|
|
│ └── 20240605_151442
|
|
│ ├── predictions
|
|
│ │ ├── internlm2-chat-1.8b-turbomind
|
|
│ │ ├── llama-3-8b-instruct-lmdeploy
|
|
│ │ └── qwen1.5-1.8b-chat-hf
|
|
│ ├── results
|
|
│ │ ├── internlm2-chat-1.8b-turbomind_judged-by--GPT-3.5-turbo-0125
|
|
│ │ ├── llama-3-8b-instruct-lmdeploy_judged-by--GPT-3.5-turbo-0125
|
|
│ │ └── qwen1.5-1.8b-chat-hf_judged-by--GPT-3.5-turbo-0125
|
|
│ └── summary
|
|
│ └── 20240605_205020 # MEMORY_SUMMARY_DIR
|
|
│ ├── judged-by--GPT-3.5-turbo-0125-charm-memory-Chinese_Anachronisms_Judgment
|
|
│ ├── judged-by--GPT-3.5-turbo-0125-charm-memory-Chinese_Movie_and_Music_Recommendation
|
|
│ ├── judged-by--GPT-3.5-turbo-0125-charm-memory-Chinese_Sport_Understanding
|
|
│ ├── judged-by--GPT-3.5-turbo-0125-charm-memory-Chinese_Time_Understanding
|
|
│ └── judged-by--GPT-3.5-turbo-0125.csv # MEMORY_SUMMARY_CSV
|
|
└── CHARM_rea
|
|
└── chat
|
|
└── 20240605_152359
|
|
├── predictions
|
|
│ ├── internlm2-chat-1.8b-turbomind
|
|
│ ├── llama-3-8b-instruct-lmdeploy
|
|
│ └── qwen1.5-1.8b-chat-hf
|
|
├── results # REASON_RESULTS_DIR
|
|
│ ├── internlm2-chat-1.8b-turbomind
|
|
│ ├── llama-3-8b-instruct-lmdeploy
|
|
│ └── qwen1.5-1.8b-chat-hf
|
|
└── summary
|
|
├── summary_20240605_205328.csv # REASON_SUMMARY_CSV
|
|
└── summary_20240605_205328.txt
|
|
```
|
|
### 3. Generate Analysis Results
|
|
```bash
|
|
cd ${path_to_CHARM_repo}
|
|
|
|
# generate Table5, Table6, Table9 and Table10 in https://arxiv.org/abs/2403.14112
|
|
PYTHONPATH=. python tools/summarize_reasoning.py ${REASON_SUMMARY_CSV}
|
|
|
|
# generate Figure3 and Figure9 in https://arxiv.org/abs/2403.14112
|
|
PYTHONPATH=. python tools/summarize_mem_rea.py ${REASON_SUMMARY_CSV} ${MEMORY_SUMMARY_CSV}
|
|
|
|
# generate Table7, Table12, Table13 and Figure11 in https://arxiv.org/abs/2403.14112
|
|
PYTHONPATH=. python tools/analyze_mem_indep_rea.py data/CHARM ${REASON_RESULTS_DIR} ${MEMORY_SUMMARY_DIR} ${MEMORY_SUMMARY_CSV}
|
|
```
|
|
|
|
## 🖊️ Citation
|
|
```bibtex
|
|
@misc{sun2024benchmarking,
|
|
title={Benchmarking Chinese Commonsense Reasoning of LLMs: From Chinese-Specifics to Reasoning-Memorization Correlations},
|
|
author={Jiaxing Sun and Weiquan Huang and Jiang Wu and Chenya Gu and Wei Li and Songyang Zhang and Hang Yan and Conghui He},
|
|
year={2024},
|
|
eprint={2403.14112},
|
|
archivePrefix={arXiv},
|
|
primaryClass={cs.CL}
|
|
}
|
|
```
|