TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

LLM-based sentiment analysis of Hacker News posts between Jan 2020 and June 2023

126 点作者 mochomocha9 个月前

21 条评论

tantalor9 个月前
Is this just using LLM to be cool? How does pure LLM with basic &quot;In the scale between 0-10 ...&quot; prompt stack up against traditional, battle-tested sentiment analysis tools?<p>Gemini suggests NLTK and spaCy<p><a href="https:&#x2F;&#x2F;www.nltk.org&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.nltk.org&#x2F;</a><p><a href="https:&#x2F;&#x2F;spacy.io&#x2F;" rel="nofollow">https:&#x2F;&#x2F;spacy.io&#x2F;</a>
评论 #41241463 未加载
评论 #41241848 未加载
评论 #41241902 未加载
visarga9 个月前
I did a similar kind of process for my own chat logs. I have about 11M tokens worth of logs, and it took 2 days to crunch all of them with ollama and LLaMA 3.1 8B on my MacBook. It&#x27;s slow, but free.<p>I generated title, summary, keywords and hierarchical topics up to 3 levels up from the original text. My plan for now is to put them in a vector search engine, which, incidentally, was made with Sonnet 3.5 with very little iteration. I want to play around to see how I can organize my ideas with LLMs, make something useful from all that text.<p>I really don&#x27;t know what I will discover. One small insight I already found is that summarization works really well, you can use summaries instead of full texts to prime Claude and it works better than expected. Unlimited context? Maybe.<p>Another direction of research is to create a nice taxonomy, there are thousands of topics, pretty difficult task, but there must be a way using clustering and LLMs. That is why I generated topic, parent-topic, gp-topic, and ggp-topic from all snippets. I would probably manually edit the top 2 levels of the taxonomy to give it the right focus.<p>I&#x27;m also integrating with my HN and reddit feeds. X is too stingy with the API. Maybe Pocket and local downloads folder too, I save&#x2F;bookmark stuff I like. I could also include all the papers I am reading into the corpus. It could synthesize a ranked feed aligned to my own interests.
评论 #41243944 未加载
评论 #41243301 未加载
huem0n9 个月前
&gt; NFL (915 posts)<p>&gt; Football (206 posts)<p>Either hacker news really likes the national forensic league, or these LLM-categories are a bit dubious.<p>Also hmmm:<p>&gt; American football (7 posts)<p>&gt; American_football (6 posts)
评论 #41248741 未加载
EarthLaunch9 个月前
&gt; Tokens Don&#x27;t Lie<p>&gt; But how do people feel about these topics<p>I find it notable that tokens don&#x27;t necessarily express people&#x27;s feelings. Put another way, tokens aren&#x27;t how people <i>feel</i>, they&#x27;re how they <i>write</i>.<p>Samstave mentioned in this thread that Twitter is a &#x27;global sentiment engine&#x27;. I&#x27;m sure that&#x27;s literally true. Sentiment measurement is only accurate to the degree that people are expressing their real feelings via tokens. I can imagine various psychological and political reasons for a discrepancy.<p>If you did sentiment analysis of publicly known writings of North Korean administrators, would that represent their feelings?<p>I think the interplay with free speech is interesting here: In a setting where people feel socially and legally safe to express their true opinion, sentiment analysis will be more accurate.
adsharma9 个月前
Can you run this tool on the removed posts dataset?<p><a href="https:&#x2F;&#x2F;github.com&#x2F;vitoplantamura&#x2F;HackerNewsRemovals">https:&#x2F;&#x2F;github.com&#x2F;vitoplantamura&#x2F;HackerNewsRemovals</a>
jmward019 个月前
I wonder if the dip is more about LLama3 70b training and data than a change in sentiment. The data cut-off was Dec 2023 for 70b. That looks to coincide with the reversal of the dip.
评论 #41241724 未加载
samstave9 个月前
&gt;&gt;<i>Use the tool below to explore various topics and the sentiments they evoke.</i><p>This is a cool phrase.<p>It is personally important as when I was asked in a panel interview @ -- They asked &quot;what do you think Twitter is?<p>My response was &quot;You&#x27;re a global sentiment engine&quot;&quot;<p>(There are a lot of conversations I&#x27;d love to have with the HN community with respect to our shared experiences, and weird history flipped-bits that exists in the minds of those who experienced that...<p>like threads of how linux came, or how xml was born through things I touched in a forrest gump way - and how there are so many stories from so many.
评论 #41241410 未加载
评论 #41241690 未加载
SubiculumCode9 个月前
I wanted to do an analysis of hacker news on another topic, but over a longer timespan.<p>I started to look into it, but in the little time I had to devote to the idea, I read that the Agolia API lets you look over a longer period, but that it is relatively costly.<p>I just want to look for all story titles from the beginning of time which match one of several simple search terms, and return submission date and title for an analysis I&#x27;d conduct in R.<p>Am I overthinking it and a simple Python script without an API code can do it?
评论 #41241759 未加载
lz4009 个月前
It&#x27;s funny filtering by crypto and seeing the (sometimes hazy) division between cryptography (we love this) and cryptocurrency (we hate it) terms.
chazeon9 个月前
I wonder if using prompts to get the sentiment in LLM is enough? So we do not need to do any fine-tuning anymore?
评论 #41242232 未加载
synicalx9 个月前
&gt; Hate : Torture<p>Great work folks, glad we can all agree on that one.<p>Interesting that they used an LLM for this. I mean it makes sense and the data seems to pass the pub test but I, in my ignorance, would not have assumed that a language model would be well suited for number crunching.
评论 #41241469 未加载
Sleaker9 个月前
Why is everything only plotted between 4 and 8 if the scale of the least liked topic should be 0 and most liked should be 9. Also 4.5 is the midpoint, but 4 is displayed as bright red and 6 is a muted gray blue, why? This makes no sense except to be psychologically disingenuous.<p>And no 5s? What is even going on in that LLM?
评论 #41242764 未加载
评论 #41243365 未加载
Mathnerd3149 个月前
&gt; Reply only the tags<p>LLM&#x27;s are really sensitive to bad or even slightly ambiguous grammar. I wonder if the numbers would differ significantly with &quot;Reply only with the tags, in the following format&quot;.
评论 #41242184 未加载
anonu9 个月前
At least Republicans and Democrats share the same low sentiment score of 4.
评论 #41242100 未加载
savin-goyal9 个月前
what&#x27;s up with the title flips from<p>&gt; 350M Tokens Don&#x27;t Lie: Love And Hate In Hacker News, to<p>&gt; LLM-based sentiment analysis of Hacker News posts, to<p>&gt; LLM-based sentiment analysis of Hacker News posts between Jan 2020 and June 2023
评论 #41242244 未加载
bravura9 个月前
Can we get a 2-d visualization of topics, and drill into topics?
评论 #41241507 未加载
elashri9 个月前
&gt; It is worth clarifying though that Hacker News does not hate International Students, but the posts related to them tend to be overwhelmingly negative, reflecting the community’s sympathy for the challenges faced by those studying abroad.<p>I was horrified when I read international students as one of top on the hate list. Although I saw a couple of comments attributed their cities housing crises on international students and thought that this sentiment is wide supported.
vtuulos9 个月前
here&#x27;s how the model ranks the discussion on this page after 40 comments:<p>SENTIMENT 6<p>:D
anonu9 个月前
Great analysis. How is divisiveness actually calculated?
评论 #41241628 未加载
评论 #41241274 未加载
thr0w9 个月前
I don&#x27;t know about this analysis and its conclusions. I&#x27;ll just use this as a jumping point to selfishly spout my own human observations.<p>For context, I&#x27;m someone who uses HN to search for topics I&#x27;m interested in, rather than something like Google or Reddit.<p>- For anything SF community-related, most hits are from 10+ years ago. Lots of &quot;hey we have a space in soma, any local startups want to hang and drink beers?&quot; or &quot;we have an empty desk in a space in the mission, any hackers want to grab it for free?&quot; - all from around 2012 or prior. Nothing like that seems to happen anymore.<p>- Starting from around 2016, a heavy anti-technology sentiment appears. Cloud, crypto, AI - all are nonsense propagated by VC types and overzealous engineers.<p>- Similarly, any thread involving money&#x2F;labor invariably has an anti-capitalist and&#x2F;or &quot;unions would solve everything&quot; tangent.<p>Would be interested to hear if others have observed similar.
评论 #41243816 未加载
teleforce9 个月前
Systemd now in the Love HN section, that a HN news in itself.