[{"data":1,"prerenderedAt":858},["ShallowReactive",2],{"project-35313171":3},{"id":4,"name":5,"line":5,"line_source":6,"title":7,"url":8,"domain":9,"maker":10,"posted_at":11,"year":12,"points":13,"comments":14,"rank_in_year":15,"shown_in_year":16,"top":17,"tier":18,"category":19,"subcategory":20,"format":21,"keywords":22,"reception":28,"sentiment":29,"stance_mix":30,"scored_comments":35,"aspects":36,"strip":42,"post_text":44,"what":45,"said":71,"maker_in_thread":141,"shape":152,"grid":157,"attempts":823,"same_week":824,"maker_shows":43,"maker_hidden":850,"reader_version":851,"card":852},"35313171","GPT-4 Reverse Turing Test","title","Show HN: GPT-4 Reverse Turing Test","https:\u002F\u002Fgist.github.com\u002Frain-1\u002F3bf56122b0ebeac929dff0f881ee8e4c","gist.github.com","rain1",1679829094,2023,288,272,160,3109,6,"legend","ai","ai evaluation experiment","other",[23,24,25,26,27],"gpt-4","turing test","reverse turing test","chatgpt","ai detection","curious",-0.181,{"positive":31,"neutral":32,"mixed":33,"critical":34},1,39,3,9,52,{"idea":37,"execution":40},{"positive":38,"critical":39},0,4,{"positive":31,"critical":41},5,[38,38,38,38,38,38,38,38,38,38,38,38,38,38,38,38,38,38,38,38,38,38,33,33,33,43,33,39,31,38,31,43,31,31,38,33,43,33,33,33],2,null,{"problem":46,"audience":44,"pricing":44,"motivation":44,"time_to_build":44,"stack":57,"openness":70,"stage":70,"team_size":70},{"value":47,"source":48,"evidence":49},"Tests whether GPT-4 can tell AI-written answers from human ones.","inferred",[50,54],{"id":51,"by":52,"text":53},"35313827","dmurray","I was hoping this would feature GPT-4 trying to distinguish between an LLM and a human pretending to be an LLM. If it were me, I'd switch focus and instead of asking \"what is…",{"id":55,"by":10,"text":56},"35314217","Oh I think I see what you mean, note that I used two different AIs (GPT-4 was the tester, Me & ChatGPT were the testee's).",[58,63],{"value":59,"source":60,"evidence":61},"GPT-4","maker",[62],{"id":55,"by":10,"text":56},{"value":64,"source":60,"evidence":65},"ChatGPT",[66,67],{"id":55,"by":10,"text":56},{"id":68,"by":10,"text":69},"35315012","I've included the exact prompt that I gave ChatGPT to get it to answer the questions without saying \"As an AI assisant\u002Flanguage model blah blah blah\"…","unknown",{"praise":72,"objections":81,"requests":108,"questions":124,"alternatives":133},[73],{"value":74,"source":75,"evidence":76},"Creative idea and fun demo","commenter",[77],{"id":78,"by":79,"text":80},"35320915","vlovich123","Haven’t LLMs been shown to be able to identify its own output vs not accurately? So this isn’t testing GPTs ability to evaluate a candidate against the Turing test so much as its…",[82,93,100],{"value":83,"source":75,"evidence":84},"Test is not interactive and differs from Turing's imitation game",[85,89],{"id":86,"by":87,"text":88},"35314715","skybrian","This is sort of interesting but it’s not interactive. Turing’s imitation game is not a written test, it’s a game, sort of like the Werewolf party game. Its difficulty is going to…",{"id":90,"by":91,"text":92},"35315741","somenameforme","The \"Turing Test\", as we talk about today, is a very simplified take on what Turing described, probably in an effort to pass it. You can read his original paper here [1]. In the…",{"value":94,"source":75,"evidence":95},"Grammar and spelling give away human answers, so the test is weak",[96],{"id":97,"by":98,"text":99},"35314581","IAmGraydon","Your grammar and spelling aren’t perfect, so that’s a dead giveaway. I wonder what the result would have been if you used perfect grammar and\u002For GPT intentionally injected some…",{"value":101,"source":75,"evidence":102},"GPT may just recognise its own output, so the test measures detection not Turing judgment",[103,104],{"id":78,"by":79,"text":80},{"id":105,"by":106,"text":107},"35315914","ec109685","Doesn’t this just boil down to whether GPT can distinguish between human written answers and GPT written ones? The actual questions don’t matter at all. It doesn’t seem like a…",[109,113],{"value":110,"source":75,"evidence":111},"Use trivia or harder questions to distinguish LLM from human",[112],{"id":51,"by":52,"text":53},{"value":114,"source":75,"evidence":115},"Tell GPT-4 the goal is passing a Turing test or use an older model",[116,120],{"id":117,"by":118,"text":119},"35314118","andybak","I'm not clear whether they asked GPT4 to pretend to be human or not? I think telling it it's goal was to pass a Turing Test would have a significant effect on it's answers.",{"id":121,"by":122,"text":123},"35313610","circuit10","Maybe try getting an older GPT-3 model that is just trained to predict text to do it rather than the new one which is trained to tell you that it’s an AI. I wouldn’t be surprised…",[125],{"value":126,"source":75,"evidence":127},"Was the tester told it was evaluating a Turing test?",[128,132],{"id":129,"by":130,"text":131},"35320800","vagab0nd","Is it just me, or is it unfair that ChatGPT doesn't know it's being tested?",{"id":117,"by":118,"text":119},[134],{"value":135,"source":75,"evidence":136},"Voight-Kampff test from Blade Runner",[137],{"id":138,"by":139,"text":140},"35313832","number6","I guess the turning test won't cut it anymore, we should use the Voight-Kampff test.",{"comments":17,"answered":39,"key":142},[143,146,149],{"id":144,"text":145},"35314143","I used to know 110 digits of pi. only know about 70 now though.",{"id":147,"text":148},"35313742","This is a great idea!",{"id":150,"text":151},"35313740","Oh! That's smart. So you could try to cut down the verbosity of the responses with a prompt, and perhaps make it harder for GPT-4 to identify. Sounds like a fun challenge!",{"commenters":153,"max_depth":154,"maker_replies":17,"first_comment_mins":155,"lifespan_hours":156},140,13,58,80,[158,162,165,168,171,175,178,181,184,187,190,192,195,198,201,204,207,210,212,215,217,219,222,225,227,230,233,236,238,241,244,247,250,252,255,258,261,264,267,270,272,274,277,279,281,283,285,287,289,291,293,296,299,301,303,305,307,310,313,316,320,323,325,328,330,333,335,337,340,343,345,348,350,352,354,357,359,361,363,365,367,369,371,373,375,378,380,383,386,389,392,395,398,401,404,406,409,412,414,416,418,421,423,425,428,431,433,436,438,440,442,444,446,448,450,452,454,457,459,461,464,466,468,471,474,476,478,480,482,484,487,490,493,496,499,501,504,506,508,510,513,515,517,519,522,524,525,527,529,532,533,536,538,540,542,545,548,550,552,555,557,560,561,564,566,567,570,574,577,579,583,587,590,591,594,596,599,602,604,607,611,614,616,618,620,622,624,628,631,633,636,639,641,643,646,649,651,654,656,660,663,666,670,674,678,680,683,685,687,691,695,699,703,705,709,712,716,719,721,723,726,729,732,734,736,739,741,743,745,747,751,755,759,761,764,767,769,771,773,775,778,781,784,786,789,792,795,797,800,802,804,806,810,812,814,816,818,820],[159,38,38,160,161,38],"35314775","other\u002Fneutral\u002Fnone","photochemsyn",[163,38,38,160,164,38],"35314908","nextworddev",[166,38,38,160,167,38],"35315045","viraptor",[169,38,38,160,170,38],"35315048","sebzim4500",[172,38,38,173,174,38],"35315648","commenter\u002Fneutral\u002Fnone","web3-is-a-scam",[176,38,38,173,177,38],"35315882","brokencode",[179,38,38,160,180,38],"35316099","moffkalast",[182,38,38,173,183,38],"35316441","IIAOPSW",[185,38,38,173,186,38],"35321508","adhesive_wombat",[188,38,38,160,189,38],"35325768","rl3",[191,38,38,173,180,38],"35326206",[193,38,38,160,194,38],"35326618","ben_w",[196,38,38,160,197,38],"35318869","jonny_eh",[199,38,38,160,200,38],"35322130","pyuser583",[202,38,38,160,203,38],"35316130","psadri",[205,38,38,160,206,38],"35316552","ilaksh",[208,38,38,160,209,38],"35317410","blatant303",[211,38,38,160,180,38],"35317859",[213,38,38,173,214,38],"35318197","LesZedCB",[216,38,38,160,180,38],"35318388",[218,38,38,160,214,38],"35322329",[220,38,38,160,221,38],"35322951","tedd4u",[223,38,38,160,224,38],"35323119","dragonwriter",[226,38,38,173,203,38],"35328310",[228,31,38,160,229,38],"35313860","isoprophlex",[231,31,38,160,232,38],"35313922","the_gipsy",[234,31,38,160,235,38],"35313963","tyingq",[237,31,38,173,229,38],"35314013",[239,31,38,160,240,38],"35313968","fullsend",[242,31,38,160,243,38],"35314023","danaris",[245,31,38,160,246,38],"35314091","TheOtherHobbes",[248,31,38,160,249,38],"35314212","prox",[251,31,38,173,243,38],"35314443",[253,31,38,173,254,38],"35314547","sho_hn",[256,31,38,160,257,38],"35315143","wizzwizz4",[259,31,38,173,260,38],"35319787","j-krieger",[262,31,38,173,263,38],"35314130","ogogmad",[265,31,38,173,266,38],"35314371","Jensson",[268,31,38,173,269,38],"35315350","mannykannot",[271,31,38,173,263,38],"35315733",[273,31,38,173,243,38],"35314452",[275,31,38,173,276,38],"35314890","rhn_mk1",[278,31,38,160,257,38],"35315162",[280,31,38,173,276,38],"35315199",[282,31,38,173,257,38],"35315282",[284,31,38,173,276,38],"35315665",[286,31,38,173,257,38],"35320977",[288,31,38,173,276,38],"35323475",[290,31,38,173,257,38],"35324226",[292,31,38,173,276,38],"35324895",[294,31,38,173,295,38],"35315774","taberiand",[297,31,38,160,298,38],"35317180","hombre_fatal",[300,31,38,160,276,38],"35317754",[302,31,38,173,229,38],"35314135",[304,31,38,173,243,38],"35314498",[306,31,38,173,269,38],"35314834",[308,31,38,173,309,38],"35315313","HervalFreire",[311,31,38,160,312,38],"35317683","dr_dshiv",[314,31,38,173,315,38],"35321436","flangola7",[317,31,38,318,319,38],"35315380","commenter\u002Fcritical\u002Fnone","kerkeslager",[321,31,38,173,322,38],"35315437","spion",[324,31,38,173,319,38],"35315539",[326,31,38,173,327,38],"35315751","baq",[329,31,38,173,319,38],"35315856",[331,31,38,173,332,38],"35317007","mynameisvlad",[334,31,38,173,319,38],"35326854",[336,31,38,173,322,38],"35332755",[338,31,38,173,339,38],"35334978","wrycoder",[341,31,38,173,342,38],"35321373","squeaky-clean",[344,31,38,173,319,38],"35326662",[346,31,38,173,347,38],"35316175","frutiger",[349,31,38,173,327,38],"35317049",[351,31,38,173,243,38],"35317633",[353,31,38,173,322,38],"35318301",[355,31,38,173,356,38],"35327083","throwaway290",[358,31,38,173,322,38],"35331256",[360,31,38,173,322,38],"35316146",[362,31,38,173,319,38],"35326579",[364,31,38,173,322,38],"35331097",[366,31,38,173,319,38],"35347002",[368,31,38,173,322,38],"35347264",[370,31,38,173,319,38],"35347489",[372,31,38,173,224,38],"35347564",[374,31,38,160,319,38],"35347165",[376,31,38,173,377,38],"35337009","zestyping",[379,31,38,173,319,38],"35346581",[381,31,38,173,382,38],"35315105","roflyear",[384,31,38,160,385,38],"35317225","cjalmeida",[387,31,38,160,388,38],"35314406","belter",[390,31,38,160,391,38],"35314435","AnIdiotOnTheNet",[393,31,38,160,394,38],"35314721","BurningFrog",[396,31,38,160,397,38],"35315326","nemo44x",[399,31,38,173,400,38],"35315788","pixl97",[402,31,38,160,403,38],"35315912","braymundo",[405,31,38,173,397,38],"35316034",[407,31,38,173,408,38],"35316600","vasco",[410,31,38,173,411,38],"35316952","cleansingfire",[413,31,38,173,397,38],"35318438",[415,31,38,160,260,38],"35319766",[417,31,38,173,397,38],"35315991",[419,31,38,173,420,38],"35316899","piloto_ciego",[422,31,38,160,397,38],"35318954",[424,31,38,173,400,38],"35320285",[426,31,38,173,427,38],"35320848","halgir",[429,31,38,160,430,38],"35323753","majkinetor",[432,31,38,173,332,38],"35316956",[434,31,38,160,435,38],"35317496","pegasus",[437,31,38,173,332,38],"35317558",[439,31,38,160,435,38],"35318724",[441,31,38,173,400,38],"35317836",[443,31,38,160,435,38],"35318638",[445,31,38,173,397,38],"35317918",[447,31,38,173,332,38],"35318275",[449,31,38,173,397,38],"35318820",[451,31,38,173,332,38],"35319397",[453,31,38,173,397,38],"35320126",[455,31,38,173,456,38],"35320150","ta8645",[458,31,38,173,332,38],"35320681",[460,31,38,173,339,38],"35334889",[462,31,38,160,463,38],"35317368","HideousKojima",[465,31,38,160,312,38],"35317634",[467,31,38,160,400,38],"35317807",[469,31,38,160,470,38],"35319244","Joeri",[472,31,38,173,473,38],"35320286","8bitchemistry",[475,31,38,173,463,38],"35321858",[477,31,38,173,473,38],"35322010",[479,31,38,173,463,38],"35322425",[481,31,38,173,473,38],"35322575",[483,31,38,173,430,38],"35323746",[485,31,38,173,486,38],"35319342","int0x2e",[488,31,38,160,489,38],"35315554","jameshart",[491,31,38,160,492,38],"35316102","shard",[494,31,38,160,495,38],"35321306","squarefoot",[497,31,38,173,498,38],"35315796","Timon3",[500,31,38,173,332,38],"35316914",[502,31,38,173,503,38],"35318033","pbhjpbhj",[505,31,38,173,332,38],"35318226",[507,31,38,160,174,38],"35315673",[509,31,38,160,206,38],"35316629",[511,31,38,160,512,38],"35316998","nextaccountic",[514,31,38,173,430,38],"35323790",[516,31,38,173,512,38],"35332227",[518,31,38,173,430,38],"35336872",[520,31,38,160,521,38],"35317874","guy98238710",[90,43,33,523,91,38],"project\u002Fneutral\u002Fidea",[51,33,33,523,52,38],[526,33,33,523,388,38],"35313874",[528,33,33,523,170,38],"35313929",[530,33,33,531,427,38],"35320961","project\u002Fneutral\u002Fexecution",[144,33,38,160,10,31],[534,33,38,160,535,38],"35314497","kimburgess",[537,33,33,531,388,38],"35315001",[539,33,33,523,170,38],"35315088",[541,33,33,531,388,38],"35315609",[543,33,33,531,544,38],"35317270","pulvinar",[546,33,33,547,388,38],"35317511","project\u002Fneutral\u002Fnone",[549,33,38,173,170,38],"35315063",[551,33,33,531,388,38],"35314634",[553,33,38,160,554,38],"35315843","esperent",[556,33,38,160,332,38],"35317073",[558,39,33,531,559,38],"35314146","imglorp",[138,41,33,523,139,38],[562,41,38,160,563,38],"35314251","Aachen",[565,41,38,160,269,38],"35314338",[121,17,33,531,122,38],[568,17,33,531,569,38],"35313646","nielsole",[571,17,31,572,573,38],"35313716","project\u002Fcritical\u002Fexecution","csomar",[575,17,38,160,576,38],"35314308","darreninthenet",[578,17,33,531,87,38],"35314964",[580,581,33,531,582,38],"35315488",7,"detuned",[584,585,33,523,586,38],"35313666",8,"imjonse",[588,585,33,523,589,38],"35313729","detrites",[147,585,38,173,10,31],[592,585,33,523,593,38],"35314022","segh",[150,585,38,595,10,31],"maker\u002Fneutral\u002Fidea",[597,585,33,531,598,38],"35314071","ModernMech",[600,585,33,531,601,38],"35313907","ape4",[603,585,33,523,563,38],"35314304",[605,585,38,160,606,38],"35314920","BuyMyBitcoins",[608,585,41,609,610,38],"35315067","project\u002Fpositive\u002Fexecution","elif",[612,585,38,173,613,38],"35317486","1970-01-01",[86,34,33,615,87,38],"project\u002Fmixed\u002Fidea",[97,617,31,572,98,38],10,[117,619,33,523,118,38],11,[621,619,31,572,563,38],"35314229",[68,619,38,623,10,31],"maker\u002Fneutral\u002Fdocs",[625,626,38,160,627,38],"35314106",12,"Alifatisk",[629,626,38,160,630,38],"35314140","namaria",[632,626,38,160,10,31],"35314188",[634,626,38,160,635,38],"35314241","p1esk",[637,626,38,173,638,38],"35323245","krainboltgreene",[640,626,38,173,635,38],"35326264",[642,626,38,173,638,38],"35361828",[644,626,38,160,645,38],"35316180","fl7305",[647,626,38,173,648,38],"35316612","qsort",[650,626,38,173,645,38],"35316931",[652,626,38,173,653,38],"35317072","phire",[655,626,38,173,648,38],"35317455",[657,626,31,658,659,38],"35319669","project\u002Fcritical\u002Fidea","cypress66",[661,626,38,160,662,38],"35318448","FranklinMaillot",[664,626,38,160,665,38],"35324392","sirsinsalot",[667,154,38,668,669,38],"35315114","other\u002Fneutral\u002Fidea","OutbackWizard",[671,672,31,572,673,38],"35317894",14,"ChancyChance",[675,676,33,523,677,38],"35313881",15,"shyamkrishna8",[679,676,33,523,170,38],"35313948",[681,676,33,523,682,38],"35313960","sd9",[55,676,38,684,10,31],"maker\u002Fneutral\u002Fexecution",[78,686,33,615,79,38],16,[688,689,31,572,690,38],"35325206",17,"darepublic",[692,693,38,160,694,38],"35316174",18,"cvsfda324",[696,697,38,160,698,38],"35320107",19,"gcanyon",[700,701,38,160,702,38],"35315378",20,"sureklix",[105,704,31,658,106,38],21,[706,707,31,658,708,38],"35319661",22,"einpoklum",[710,711,38,160,610,38],"35314951",23,[713,714,38,160,715,38],"35313921",24,"1equalsequals1",[717,714,38,173,718,38],"35314031","FrustratedMonky",[720,714,38,173,391,38],"35314529",[722,714,38,160,665,38],"35324406",[724,714,38,160,725,38],"35315084","Frost1x",[727,714,38,160,728,38],"35315617","slfnflctd",[730,714,38,160,731,38],"35319716","sysstem",[733,714,38,160,665,38],"35324403",[735,714,38,173,563,38],"35314316",[737,714,38,160,738,38],"35314454","mjburgess",[740,714,38,173,662,38],"35318513",[742,714,38,173,715,38],"35318800",[744,714,38,173,662,38],"35323848",[129,746,33,523,130,38],25,[748,749,33,523,750,38],"35316129",26,"ouid",[752,753,33,547,754,38],"35318715",27,"rendall",[756,757,33,523,758,38],"35313695",28,"mlcrypto",[760,757,38,160,170,38],"35313744",[762,757,38,160,763,38],"35313911","Trasmatta",[765,757,31,658,766,38],"35313751","capableweb",[768,757,33,523,388,38],"35313904",[770,757,33,523,766,38],"35314532",[772,757,33,523,388,38],"35314790",[774,757,33,523,766,38],"35316794",[776,757,38,160,777,38],"35317357","Takennickname",[779,757,38,160,780,38],"35313761","beefield",[782,757,38,160,783,38],"35313839","drstewart",[785,757,38,160,563,38],"35314345",[787,757,38,173,788,38],"35314768","LudwigNagasena",[790,757,38,173,791,38],"35313791","dbingham",[793,757,38,173,794,38],"35313910","aka878",[796,757,38,160,315,38],"35321530",[798,757,38,173,799,38],"35313977","qingdao99",[801,757,38,173,598,38],"35314040",[803,757,38,173,799,38],"35314080",[805,757,33,547,170,38],"35315180",[807,757,33,808,809,38],"35314026","project\u002Fmixed\u002Fnone","WastingMyTime89",[811,757,33,547,563,38],"35314367",[813,757,33,547,266,38],"35314566",[815,757,33,547,809,38],"35315689",[817,757,38,173,194,38],"35326563",[819,757,38,173,777,38],"35317316",[821,757,38,160,822,38],"35314219","9991",[],[825,830,834,838,842,846],{"id":826,"name":827,"points":828,"category":829},"35316679","Apple Notes Liberator",563,"productivity",{"id":831,"name":832,"points":833,"category":19},"35334848","Regex.ai",294,{"id":835,"name":836,"points":837,"category":19},"35350662","Libraria",271,{"id":839,"name":840,"points":841,"category":829},"35327112","Open AI is not Open",241,{"id":843,"name":844,"points":845,"category":19},"35301657","llm-security",238,{"id":847,"name":848,"points":849,"category":19},"35344787","Lit-LLaMA",158,false,"v2",{"id":853,"name":5,"line":5,"ai":850,"cat":19,"year":12,"points":13,"comments":14,"rec":28,"s":854,"top":17,"kw":855,"maker":10,"domain":9,"tier":18,"attempts":38,"strip":856,"url":8,"mix":857,"praise":74,"objection":83,"sub":20},35313171,-0.18,[23,24,25,26,27],"0000000000000000000000333234101211032333",[31,32,33,34],1794036633061]