The code of our paper "InfLLM: Unveiling the Intrinsic Capacity of LLMs for Understanding Extremely Long Sequences with Training-Free Memory"
Geek Repo:Geek Repo
Github PK Tool:Github PK Tool
cat-sun opened this issue 3 months ago · comments
请问调用bash infinitebench.sh,在longbook_choice_eng数据集上测了50条数据,准确率只有40%左右,这个结果是不是有问题,是中间过程出了什么问题吗
您好,longbook_choice_eng数据上,InfLLM的表现为43.23,因此您的测试应该是准确的