Analyzing top HN posts with language models
Data
text clustering analysis
SHOW HN2022
Analyzing top HN posts with language models
Analyzing top HN posts with language models
news.ycombinator.com#hacker news#text embeddings#clustering#umap#language models
How HN took it
4 positive, 2 neutral or asking, 0 mixed, 5 critical, out of 11 comments about the project. Thread score -0.10.
Execution: 2 for, 2 againstIdea: 0 for, 2 againstDesign: 0 for, 1 against
The thread, comment by comment
Each row is a top-level comment and its replies, in HN's order. Select a cell to open that comment on HN.
positive neutral or mixed critical about another commenter or something else the maker
What HN said
What people liked
Interesting analysis and useful visualizations
From 3 comments
Nice idea and analysis! I reproduced it as well with https://graphext.com and got similar clusters…
victorianoi on HNI love this shit, dude. Not for any great insight. I just enjoy this sort of thing. Good stuff.
renewiltord on HNLove your blog posts and visualizations Jay, thanks for sharing!
arolihas on HN
Embeddings and data shared freely for reuse
From 1 comment
Thanks. I just added a disclosure to the comment (can't edit the parent anymore). The full embeddings are freely provided here without the need to use the service.
jayalammar on HN
Objections
Undisclosed conflict of interest: author works at Cohere, which sells the embedding service
From 3 comments
The conflict of interest here concerns me. I don't object to content marketing, but I'd rather a) you were clear from the start that you work for this company and are promoting…
wpietri on HN"Normal" doesn't mean good, so that line of argument doesn't work for me. But the specific "stink" I'm objecting to here is, as I said, a conflict of interest. In specific,…
wpietri on HNand the more fake reviews, the less I trust them. Likewise, the more people here advertising their products without proper disclosure, the less I'll trust the honesty of the…
woojoo666 on HN
Clusters look dense and poorly separated in the UMAP projection
From 1 comment
Interesting, but it doesn't seem like the dimensionality reduction produces a good separation of topics. The UMAP projection looks pretty dense. Did you consider pruning or using…
uniqueuid on HN
Recency bias in the dataset because top post counts and scores grew over time
From 1 comment
BTW there is an implicit recency bias in the dataset, since 2017 the number of top 3K post became more frequently and the avg score is larger year after year as the community in…
victorianoi on HN
Ask HN posts are naturally biased toward engagement
From 1 comment
I don't know how HN score metrics work but after some short review of the datafile [1] I've noticed a lot of the posts has the form of a simple questions and as such seems to be…
toppy on HN
Requests
Try t-SNE or other clustering approaches for better separation
From 2 comments
Try t-SNE. I used to scoff at cluster plots until I saw those but with t-SNE… wow, those clusters are actually separated!
PaulHoule on HNInteresting, but it doesn't seem like the dimensionality reduction produces a good separation of topics. The UMAP projection looks pretty dense. Did you consider pruning or using…
uniqueuid on HN
Add additional engagement metrics to remove bias
From 1 comment
I don't know how HN score metrics work but after some short review of the datafile [1] I've noticed a lot of the posts has the form of a simple questions and as such seems to be…
toppy on HN
Questions
Why cluster on original embeddings rather than UMAP output
From 2 comments
Interesting, but it doesn't seem like the dimensionality reduction produces a good separation of topics. The UMAP projection looks pretty dense. Did you consider pruning or using…
uniqueuid on HNSo it really depends on what you use for clustering. In this case, I'm clustering by the original embeddings so the UMAP results are different. I've also seen: 1- Clustering by…
jayalammar on HN
Is top score a sign of quality or just early posting
From 1 comment
There's too much fixation with "top" in our industry. Top voted tends to mostly be a function of early posting. Later posts don't get votes because they simply were not seen.…
natch on HN
Alternatives named
graphext
From 1 comment
Nice idea and analysis! I reproduced it as well with https://graphext.com and got similar clusters…
victorianoi on HN
Ploomber HN classifier
From 1 comment
I grouped posts by topic: https://n3ws.ploomber.io Explanation: https://ploomber.io/blog/hn_classifier/
hoerzu on HN