Hugging Face's logo

Datasets:
AnonymousSite
/
QA_test_dataset_for_CEM_MCQs

Modalities:
Text
Formats:
csv
Size:
1K - 10K
Libraries:
Datasets
pandas
License:
Dataset card Viewer Files Community
1

You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

1. Introduction

(1) The CEM-QA test dataset comprises 2,825 questions from Registered Constructor Qualification Examinations (RCQEs) and 2,225 experts-designed questions covering 7 second-tier subdomains.

(2) Each question from the CEM-QA test dataset are manually annotated with four features, including the question source, single-answer multiple-choice question (SAMCQ) or multiple-answer multiple-choice question (MAMCQ), and calculation question or non-calculation question.

(3) This CEM-QA test dataset is developed and maintained by Southeast University, University of Cambridge, China University of Mining and Technology, Nanjing University of Aeronautics and Astronautics, and University of Macau.

(4) Make sure to read the specification and follow the rules.

2. Submission of your GLLM’s answers

The answers could be submitted through https://forms.gle/v7EhLwyc3uRmxfJLA. Please use “Template of answer submission.xls” in this repository to submit your GLLM's answers.

3. Citation requirement

The reuse of this repository requires citation. Should any individual or entity utilize this repository without appropriate acknowledgment and citation, they do not have the right to use our data. We will take measures to protect our copyright, including, but not limited to, retracting their papers and initiating legal action.

4.GLLM Leaderboard for CEM-QA

General-purpose large language models Contributors Average correctness ratio SD1 SD2 SD3 SD4 SD5 SD6 SD7 Ranking
Knowledge-incorporated ERNIE-Bot 4.0 Baidu & The authors 0.734 0.787 0.758 0.640 0.766 0.784 0.754 0.808 1
Knowledge-incorporated fine-tuned Qwen-14B-Chat Alibaba & The authors 0.631 0.683 0.618 0.536 0.659 0.683 0.629 0.719 2
Knowledge-incorporated GPT-4 OpenAI & The authors 0.620 0.672 0.611 0.431 0.694 0.695 0.694 0.678 3
Original ERNIE-Bot 4.0 Baidu 0.608 0.661 0.632 0.507 0.656 0.648 0.632 0.622 4
Knowledge-incorporated Qwen-14B-Chat Alibaba & The authors 0.583 0.641 0.565 0.465 0.621 0.641 0.578 0.669 5
Original Qwen-14B-Chat Alibaba 0.522 0.583 0.512 0.393 0.565 0.548 0.521 0.594 6
Knowledge-incorporated fine-tuned Baichuan2-7B-Chat Baichuan AI & The authors 0.517 0.568 0.701 0.224 0.682 0.682 0.682 0.682 7
Knowledge-incorporated fine-tuned Qwen-7B-Chat Alibaba & The authors 0.517 0.574 0.538 0.410 0.567 0.575 0.514 0.597 8
Knowledge-incorporated fine-tuned Baichuan2-13B-Chat Baichuan AI & The authors 0.497 0.547 0.491 0.413 0.566 0.552 0.494 0.553 9
Original GPT-4 OpenAI 0.475 0.513 0.477 0.358 0.512 0.480 0.528 0.486 10
Knowledge-incorporated Qwen-7B-Chat Alibaba & The authors 0.468 0.529 0.448 0.364 0.515 0.514 0.467 0.542 11
Knowledge-incorporated fine-tuned GPT-3.5-turbo OpenAI & The authors 0.468 0.497 0.441 0.393 0.513 0.497 0.498 0.594 12
Knowledge-incorporated Baichuan2-7B-Chat Baichuan AI & The authors 0.444 0.484 0.474 0.366 0.495 0.474 0.417 0.489 13
Knowledge-incorporated Baichuan2-13B-Chat Baichuan AI & The authors 0.441 0.479 0.429 0.371 0.502 0.481 0.430 0.514 14
Knowledge-incorporated fine-tuned ERNIE-Bot-turbo Baidu & The authors 0.427 0.487 0.401 0.374 0.479 0.487 0.436 0.583 15
Knowledge-incorporated fine-tuned ChatGLM3-6B Tsinghua & The authors 0.425 0.482 0.406 0.353 0.471 0.487 0.422 0.472 16
Original Qwen-7B-Chat Alibaba 0.410 0.461 0.370 0.316 0.475 0.445 0.423 0.411 17
Knowledge-incorporated GPT-3.5-turbo OpenAI & The authors 0.407 0.458 0.395 0.356 0.438 0.456 0.422 0.447 18
Knowledge-incorporated ChatGLM3-6B Tsinghua & The authors 0.399 0.462 0.395 0.314 0.454 0.452 0.394 0.406 19
Original Baichuan2-13B-Chat Baichuan AI 0.393 0.443 0.408 0.323 0.447 0.443 0.378 0.456 20
Knowledge-incorporated ERNIE-Bot-turbo Baidu & The authors 0.392 0.424 0.386 0.351 0.432 0.418 0.394 0.467 21
Original Baichuan2-7B-Chat Baichuan AI 0.385 0.423 0.406 0.291 0.445 0.427 0.381 0.394 22
Original ChatGLM3-6B Tsinghua 0.353 0.411 0.351 0.298 0.394 0.403 0.343 0.339 23
Original ERNIE-Bot-turbo Baidu 0.345 0.402 0.309 0.324 0.382 0.370 0.365 0.414 24
Original GPT-3.5-turbo OpenAI 0.340 0.400 0.334 0.304 0.421 0.345 0.362 0.389 25
Downloads last month
19