ai检测真是呵呵了

However, the application of emerging technologies in new disciplines often
encounters challenges due to subjective biases and other factors. For instance, there
may be an overemphasis on the application of certain software, leading to a failure
to correctly interpret bibliometric indicators in studies and to conduct objective and
comprehensive analyses of the research feld. Therefore, it is necessary to critically
evaluate and assess existing research on the application of bibliometric mapping technology in educational technology. By understanding the current state of research, we
can assess the methodological system’s level of maturity, the organizational structure
of the content, and the credibility of research conclusions in the feld. Considering the
limitations observed in prior studies involving the utilization of bibliometric mapping
techniques and the absence of comprehensive reviews concerning these applied methodologies in the feld of Educational Technology, this systematic review employs a
methodology akin to that of previous meta-reviews (referred to as reviews of reviews)
(i.e. Polanin et al., 2016; Tamim et al., 2011). Its primary objective is to function as a
meta-review of existing bibliometric studies. The intention is to uncover defciencies,
discrepancies, disagreements, and incongruities within bibliometric studies, pinpoint
fundamental elements that should be integrated into standard bibliometric investigations, and provide valuable insights to steer future development。这段来自已经发表的文章,


我去测试ai率,zerogpt和phrasely都是0%ai(这两个我怀疑是一个测试模式,每次都是一样的),stealthwriter是33%,quillbot是100%,哈哈哈哈,太好笑了!你以为quillbot最严吗?其实不好说,最起码我拿我的雅思作文去测试的时候,quillbot显示0%ai,结果stealthwriter显示50%ai…… 我看很多人在研究降低ai率,我觉得那些学校和期刊能不能自己内部整明白了?你们特么的到底要信哪家的ai检测结果???PS: 有大佬知道什么靠谱的降低ai方法吗?我怎么感觉降低ai的方法就是让文本变得更加小学生呢?好家伙,我多用高级词汇、多用复合句、多用首先其次什么的、写作时候句子结构比较统一……这些都是ai特征?我学习这么多年,不就是要学会这种写法吗?你特么跟我说这是ai特征???我降低ai,就得用低级词汇、减少从句、不用连词???我雅思那些东西都白学了呗?

4 个赞

学了这么多年,我成替身了

我自己用过stealth writer的改写功能,默认10分(人性化程度),改出来一看——这什么玩意?这不是雅思5.5-6的垃圾水平吗?这玩意能用在论文里面?无语

我觉得挺正常的。。

我有个朋友申请大学自己写的问题,结果turnitin AI检测率快到80%(还是70%来着)了 :tieba_025:

准确性差远了,误判几率太高了。越是严谨、有逻辑性、语言水平高,越是误判,你写的内容越是口语化、语无伦次,这样ai率越低,真的麻了。发明llm写东西,然后又各种检测ai率,在学术界和seo压制llm生成的内容。那llm干脆封禁写作功能好了……

越是严谨、有逻辑性、语言水平高,越是误判

对对对,高级词汇直接就寄了。

:tieba_025:这个我之前研究过,其实就是Ai套皮+一些词汇进行+分,然后最终得分。怎么说呢,如果你的内容是GPT-3.5生成的,那么用GPT-4去检测+词汇+分,那么识别率很高。但是,如果你的内容是GPT-4生成的,或者级别比检测的Ai更高,检测得分是人工写的。

:tieba_025:一句话:我的Ai在你之上,你无法检测判断

stealth writer貌似检测模型蛮多的,尤其是单独列出来了Claude,其他模型可能还在盯着gpt4呢,所以有的人用Claude写了去测试就没事。但是stealth writer 反正无论是Claude还是纯人工都能测试出来ai比例,我也是服了。用了它的改写,雅思直接白学的感觉,改的什么小学生英语?但是去测试几个ai检测网站,的确都能通过,真是哭笑不得,ai比例是降低了,英语水平也没法看了…… 感觉llm进化到最后应该是无法检测的,感觉目前是人类在垂死挣扎

正是因为这种文风优秀所以AI才模仿了,我觉得只要保证写作质量足够好就行了,没必要管是不是AI生成的。那种低质量的AI文章的本质问题也是低质而非AI

越是专业性高的,语术严谨的,他越会认为你是ai,相反,错别字多的,表单符号有问题的,他会认为你是正常人

From #develop:ai to 资源荟萃