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.

ArxivGPT: Chrome extension that summarizes arxived research papers using ChatGPT

154 pointsby excerionsforteover 2 years ago

20 comments

goldemeraldover 2 years ago
The example here is a bit worrying for the peer review process. I am not looking forward to my &quot;peers&quot; reviewing my paper by putting it through LLMs and blindly copy pasting the output. I can already imagine emailing the Area Chair and saying &quot;While reviewer 2 is detailed, the questions show a severe lack of basic understanding. We believe the contents are AI generated.&quot;<p>Then again, perhaps LLMs could simply be incorporated into the peer-review process, where after submitting your paper, you&#x27;d have to answer the AI&#x27;s basic questions. As a reviewer, I could imagine a structured AI report for a paper being helpful in guiding discussion: &quot;The paper compares to recent approaches X, Y, and Z. And the work is indeed novel.&quot;
评论 #34770916 未加载
评论 #34771572 未加载
评论 #34771761 未加载
irthomasthomasover 2 years ago
Caution. Language models do not know what is salient to a human. They also have a strong bias toward information that they have seen frequently. Research will contain a larger amount of new information and it&#x27;s that new information which is most valuable to us, but least relevant according to the models.
评论 #34773411 未加载
herbstover 2 years ago
Ive used ChatGPT to make short factual videos for YouTube and honestly it&#x27;s a bit worrying with supposed &#x27;facts&#x27;<p>I would not suggest anyone to use ChatGPT outputs for actual knowledge at this point.
评论 #34771358 未加载
评论 #34771671 未加载
评论 #34772488 未加载
评论 #34771382 未加载
评论 #34772185 未加载
juliangmpover 2 years ago
Man if only the authors of such papers would write small summaries about the content and put them in the papers themselves or something
评论 #34771590 未加载
评论 #34772205 未加载
m3kw9over 2 years ago
This is dangerous because people that has no knowledge of the science would blindly trust whatever it summarizes. There is no way to verify, an example is if you ask to summarize a book that you have understanding of the subject, at least you can sense some bs or open up the book at verify a few points. Here you would be at GPT3 mercy
评论 #34775572 未加载
评论 #34775654 未加载
评论 #34775200 未加载
wonderfulyover 2 years ago
Great work! Glad to see innovations based on my work <a href="https:&#x2F;&#x2F;github.com&#x2F;wong2&#x2F;chatgpt-google-extension">https:&#x2F;&#x2F;github.com&#x2F;wong2&#x2F;chatgpt-google-extension</a> That&#x27;s why I open sourced the code!
ggmover 2 years ago
Don&#x27;t trust the output to reflect what&#x27;s in the paper.
评论 #34771765 未加载
评论 #34772753 未加载
voxelghostover 2 years ago
For my paper, it just straight up invented most of the &quot;relevant references&quot; section
aix1over 2 years ago
Isn&#x27;t this what abstracts are for?
评论 #34770797 未加载
评论 #34771217 未加载
评论 #34770780 未加载
JangoSteveover 2 years ago
I think my favorite part of this prompt is that it starts with, &quot;Please...&quot;<p>With this new class of products based on crafting prompts that best exploit a GPT&#x27;s algorithm and training data, are we going to start seeing pull requests that tweak individual parts or words of the prompt. I&#x27;m also curious how the test suite for projects like this would look for specific facts or phrases to be contained in the responses for specific inputs.
iamflimflam1over 2 years ago
Would love something that would translate papers from academic speech into something I can enjoy reading.
评论 #34771176 未加载
newswasboringover 2 years ago
To make it work in brave we need to turn off language based finger printing [1]. I wonder how&#x27;s that related.<p>Edit: btw, congratulations on the release. This is the kind of stuff I think should be explored more using LLMs. Great choice on making a chrome extension, it&#x27;s great UI for this kind of thing.<p>[1] <a href="https:&#x2F;&#x2F;github.com&#x2F;hunkimForks&#x2F;chatgpt-arxiv-extension#how-to-make-it-work-in-brave">https:&#x2F;&#x2F;github.com&#x2F;hunkimForks&#x2F;chatgpt-arxiv-extension#how-t...</a>
inductive_magicover 2 years ago
This isn&#x27;t viable because of bias and blatant lies in LLM-outputs.<p><a href="https:&#x2F;&#x2F;huggingface.co&#x2F;ml6team&#x2F;keyphrase-extraction-kbir-inspec" rel="nofollow">https:&#x2F;&#x2F;huggingface.co&#x2F;ml6team&#x2F;keyphrase-extraction-kbir-ins...</a> is a decent tool to explore the constant stream of publications. The last mile still is left to the human.
alexanderdavideover 2 years ago
I&#x27;m wondering about two things related to development of customer-facing programs that use paid APIs in the background:<p>Are you using your own API key and pay for the usage? How can you justify operation of programs that produce high costs but no income? Isn&#x27;t the API publicly exposed to the client-side and possible subject of theft and abuse?
评论 #34771548 未加载
评论 #34771550 未加载
meghan_rainover 2 years ago
What if the paper is longer than what fits in a prompt?
评论 #34771161 未加载
评论 #34773741 未加载
f_allweinover 2 years ago
I don’t understand. Don’t research papers usually have a summary, provided by the authors?
seaucreover 2 years ago
The related references don&#x27;t appear to be real.
2Gkashmiriover 2 years ago
firefox extension dude
amitportover 2 years ago
Also works on Edge
dankleover 2 years ago
Is it good?