OpenLMLab/GAOKAO-Bench
GAOKAO-Bench is an evaluation framework that utilizes GAOKAO questions as a dataset to evaluate large language models.
PythonApache-2.0
Issues
- 1
是否支持其他开源模型本地调用,不是以apikey方式
#35 opened by jidandan666 - 3
- 1
Missing Question
#30 opened by MarigoldTechStriker - 1
- 1
Solution of "Exception: list index out of range" when run `python choice_bench.py`
#18 opened by ALLinLLM - 3
json. decoder.JSONDecodeError: Invalid escape:
#19 opened by zyy-2001 - 1
Typos in the dataset.
#20 opened by chengeharrison - 1
关于测试题目格式
#21 opened by jimmyzhang610 - 1
gpt-3.5-turbo 高考得分 在哪?
#7 opened by bansky-cl - 2
有GPT-4的具体实验结果吗
#28 opened by K1yomi - 4
请问1000道主观题都是完全的人工评分吗?涉及到用gpt4来评分吗?
#23 opened by eyuansu62 - 0
按照README文档运行Openai Api简单实例,运行choice_bench.py后输出2010-2022_Math_II_MCQs single_choice mkdir: cannot create directory ‘../data/Multiple-choice_Questions/gpt-3.5-turbo_2010-2022_Math_II_MCQs’: File exists 0%| | 0/44 [00:00<?, ?it/s]Exception: Cannot choose from an empty sequence 0%| | 0/44 [00:00<?, ?it/s]Exception: Cannot choose from an empty sequence 0%| | 0/42 [00:00<?, ?it/s]Exception: Cannot choose from an empty sequence 0%| | 0/44 [00:00<?, ?it/s]Exception: Cannot choose from an empty sequence 0%| | 0/44 [00:00<?, ?it/s]Exception: Cannot choose from an empty sequence Exception: Cannot choose from an empty sequence
#27 opened by FreshOrangess - 1
请问`Gaokao-2023`这个数据集去哪里找呢?
#26 opened by zhimin-z - 2
请问关于评估过程是0shot评测还是5shot评测?
#24 opened by liu904-61 - 0
这个评测怎么没有排除主观评测
#22 opened by eyuansu62 - 2
有没有gpt4的得分?
#4 opened by vsEcho567 - 2
- 1
- 2
- 0
- 1
Bench 目录下三个 prompt.json 文件中有 <eoa> 错写成 <eoe>
#6 opened by Leymore