电脑桌面
添加运营动脉到电脑桌面
安装后可以在桌面快捷访问

大型语言模型对学习认知负载的影响分析会员免费

大型语言模型对学习认知负载的影响分析_第1页
1/206
大型语言模型对学习认知负载的影响分析_第2页
2/206
大型语言模型对学习认知负载的影响分析_第3页
3/206
大型语言模型对学习认知负载的影响分析_第4页
4/206
大型语言模型对学习认知负载的影响分析_第5页
5/206
大型语言模型对学习认知负载的影响分析_第6页
6/206
大型语言模型对学习认知负载的影响分析_第7页
7/206
大型语言模型对学习认知负载的影响分析_第8页
8/206
大型语言模型对学习认知负载的影响分析_第9页
9/206
大型语言模型对学习认知负载的影响分析_第10页
10/206
Your Brain on ChatGPT: Accumulationof Cognitive Debt when Using an AlAssistant for Essay Wrting TaskNataliyaKosmynaEugeneHauptmannYeTongYuanJessicaSituMITMediaLabMITWellesleyCollegeMITCambridge,MACambridge, MAWellesley,MACambridge,MAXian-HaoLiaoAshlyVivianBeresnitzkyIris BraunsteinPattieMaesMass.CollegeofArtMITMITMediaLabandDesign(MassArt)Cambridge, MACambridge,MACambridge,MABoston,MAUnitedStatespvaluesEEGAlphaBandFigure1.ThedynamicDirectTransterFunction(dDTF)EEGanalysisofAlphaBandforarounsLLMSearchEnaine.Brain-only,includingp-valuesto show siginificancefrommoderatelysignificant(")tohighlysignificant(**,NataliyaKosmynais thecorresponding author, please contact herat nkosmvna@mit edluDistributed underCCBY-NC-SA

AbstractWith today's wide adotion of LLM products like ChatGPT fromOpenAl, humans andbusinesses engage and use LLMs on a daily basis. Like any other tool, it carries its own set oadvantages andlimitations. Thisstudyfocusesonfindina outthecoanitive cost ofusina an !!Min the educational context of writing an essav.a We assianed participants to thee groups:ILM group,Search Engie group, Brain-only groupwhere each participant used a desianated tool (or no tool in the latter) to write an essay Weconducted 3 sessions with the same aroup assinment for each particiant.In the 4th sessioweaskedLLMgroupparticipants touse no tools (we refer to them as LLM-to-Brain), and theLM). We recruited a totalof 54participantsforSessions1,2,3,and18particiantsamonathemcomoletedsession4Weusedelectroencephalographv(EEG) to record participants'brain activity in order to assessnitive load. and to aain adeeper understandina ofneuraactivations during the essay writina task. We performed NLP analysis. and we interviewed eachand an Aliudae (a specially built Al agent)EntitiesRecoanition(NERs)LLM. SearchEnaineandBrain-onlv aroushadsianificantlvdifferentneuralconnectivityBrainconnectivitv svstematicallyscaled downwiththe amount of externalsupport:the Brain-onlv aroup exhibited the stronaest. widest-ranginenetworks,SearchEnginegroupshowedintermediateenaaaementandLLMassistanceelicitethe weakestoverall couplina. Insession participants showed weaker neuraandunder-engagemenparticipantsdemonstratedhigherwidespreadoccipito-parietal and prefrontal nodethe visual processina. similar to the onefrequentlvperceivedintheSearchThe reoorted ownership of LLM group'sessays in the interviews was ow The Search Engine group had strong ownership, bur lessethan the Brain-only group. The LM group also fel behind in their ability to qucte fom theessavs thev wrote just minutes prionAs the educationalimpact ofL.Museonly beains to setle with the aeneral oooulation. in thisstudv we demonstrate the pressina matter of a likela learning skills based on theresultsofourstudy.Theuseof LLMhad ameasurable impact on oarticipants. and while thebenefitswereinitiallyapparent,aswedemonstratedoverthecourseof4months.theLLMgroup'sparticipantsperformedworsethantheircounterpartsintheBrain-onlyaroupatallevelsneural, linguistic, scoringWe hope this study serves as a preliminary...

1、当您付费下载文档后,您只拥有了使用权限,并不意味着购买了版权,文档只能用于自身使用,不得用于其他商业用途(如 [转卖]进行直接盈利或[编辑后售卖]进行间接盈利)。
2、本站所有内容均由合作方或网友上传,本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺!文档内容仅供研究参考,付费前请自行鉴别。
3、如文档内容存在违规,或者侵犯商业秘密、侵犯著作权等,请点击“违规举报”。

查找下载文件的指引

一、电脑端

- Windows 系统:按下键盘快捷键 `Ctrl + J`,即可打开下载列表。  

- Mac 系统:按下键盘快捷键 `⌘ + J`,即可打开下载列表。  

二、手机端

1. 打开手机浏览器,点击浏览器右下角的 “≡”(或“更多”)图标。  

2. 在弹出的菜单中找到并点击 “下载内容”(或类似选项),即可查看已下载的文件。  

提示:不同浏览器界面略有差异,若未找到“下载”入口,可尝试在浏览器设置中搜索“下载”关键词。


大型语言模型对学习认知负载的影响分析

确认删除?
会员
教程
收藏
足迹
联系
  • 站长微信
回到顶部