TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Show HN: I trained GPT-2 to write like Goop

96 pointsby calebkaiserabout 5 years ago

9 comments

c3534labout 5 years ago
After trying a few to get a handle on things, I decided to make guesses based solely on if the sentence makes coherent sense. I judged 14 quotes before getting a repeat, having gotten 13 wrong and only 1 right.<p>Now, if it was 50&#x2F;50, I&#x27;d have said they&#x27;re just as coherent. But it&#x27;s 13:1 which suggests to me that there is a bias here. I think the authors intentionally selected quotes which make the least sense out of context to be the Goop quotes and cherry picked GPT2 quotes that happened to sound the most sense without any context. This is supported by the fact that I only had to go through 14 quotes before it started repeating.<p>If that&#x27;s the case, and I suspect it is, it&#x27;s not really dishonest per se, but it is at least sensationalist and potentially misleading. It&#x27;s asking you to draw conclusions by having you participate in an experiment where it has its thumb on the scales.
评论 #23076880 未加载
评论 #23078654 未加载
评论 #23077729 未加载
评论 #23079276 未加载
评论 #23079181 未加载
calebkaiserabout 5 years ago
Hey HN -- I had a slow weekend and built this thing. It&#x27;s a little game that displays a sentence, and you decide if you think an ML model generated it, or if it&#x27;s an actual quote from Goop.<p>I fine tuned the model (OpenAI&#x27;s GPT-2) using Max Woolf&#x27;s gpt-2-simple and by scraping articles from Goop&#x27;s &quot;Wellness&quot; section. I generated predictions by feeding it a few words from the opening of actual Goop sentences (not sentences it was trained on) and seeing what it spat out.<p>There aren&#x27;t many quotes (something like 25) in it right now, but I can add more easily if people have fun with it.
评论 #23080151 未加载
评论 #23076575 未加载
elil17about 5 years ago
I can&#x27;t believe how bad I am with this! It reminds me of the book &quot;The Most Human Human.&quot; Every year, at the competition for the Loebner Prize (a five minute chatbot Turing test), an award is given not only to the AI which convinces the most humans that it is a human (The Most Human Machine), but also to the human who convinces the most humans that they are a human (The Most Human Human). The author of the book tried to set the record for being the highest scoring human in the test. His book dives into just how smart, curious, and empathetic you have to be to show your human-ness. The Goop authors, it seems, are not going to be in the running for &quot;Most Human Humans&quot; any time soon.
评论 #23076550 未加载
mattigamesabout 5 years ago
For people like me who didn&#x27;t know: Goop.com describes itself as &quot;Cutting-edge wellness advice from doctors, vetted travel recommendations, and a curated shop of clean beauty, fashion, and home.&quot; The company has been the eye of many criticisms specially for lack of scientific prove of their healthcare advices <a href="https:&#x2F;&#x2F;en.m.wikipedia.org&#x2F;wiki&#x2F;Goop_(company)" rel="nofollow">https:&#x2F;&#x2F;en.m.wikipedia.org&#x2F;wiki&#x2F;Goop_(company)</a>
评论 #23076654 未加载
评论 #23079265 未加载
andy_pppabout 5 years ago
When I see things like this I start to wonder if all of deep learning is actually getting this close to an uncanny valley like result, but because it doesn&#x27;t really understand grammar (or an image, or a road etc.) we don&#x27;t feel the weirdness of the result. There is just something a bit off about the structure and meaning of what is produced such that reading the ones not written by Goop usually clog up my brain. The Goop ones clog up my brain in a different way of course ;-)
calebkaiserabout 5 years ago
Hey everyone—looks like HN traffic squeezed my server a little more than I expected. It&#x27;s upgrading now, but might not be available for about 5 minutes. Sorry about that, and thanks for checking it out!<p>EDIT: All good now :)
评论 #23078093 未加载
martopixabout 5 years ago
Very similar: <a href="http:&#x2F;&#x2F;snarxiv.org&#x2F;vs-arxiv&#x2F;" rel="nofollow">http:&#x2F;&#x2F;snarxiv.org&#x2F;vs-arxiv&#x2F;</a>
DrScumpabout 5 years ago
This reminds me of the time that MAD Magazine did parodies of entries in the old Spencer Gifts catalogs.
fake-nameabout 5 years ago
I mean, it can&#x27;t have been too hard. Anyone can do this. Just open your laptop and take a big watery shit right on the keyboard.<p>Volia, Content indistinguishably coherent from the original. Bring your own sanitation wipes.