Inspectus
AI
model interpretability visualization
SHOW HN2024
197points
15comments
#400of 3,515 in 2024
Top 12%of 2024 by points
How HN took it
CuriousMostly questions and neutral comments
2 positive, 7 neutral or asking, 0 mixed, 1 critical, out of 10 comments about the project. Thread score +0.10.
Design: 1 for, 0 againstDocs: 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
Useful for understanding model behavior and debugging
From 2 comments
Hey! This is pretty neat, it reminds me of the graphs made by transformer_lens. Cool to see all of these visualization libraries popping up!
JackYoustra on HNI'm not a primary user. Just cleaned up the existing codebase to make it open source. But you could use this to visualise attentions and debug the model. For an example if you're…
lakshith-403 on HN
Objections
Documentation and usage guidance are lacking
From 2 comments
The issue with the ambiguity of usage plague lots of OSS projects. Guides/Tutorials will always help drive usage much more, just look at the usage of GPT-3 vs ChatGPT (which is…
3abiton on HNThis looks cool but can you explain how to make it useful?
benf76 on HN
Requests
Walkthrough tutorial or guide for non-engineers
From 2 comments
Sounds great. Non-engineer, but curious. Is there a walkthrough blog post or video that can help someone appreciate/understand this easily?
ravjo on HNThe issue with the ambiguity of usage plague lots of OSS projects. Guides/Tutorials will always help drive usage much more, just look at the usage of GPT-3 vs ChatGPT (which is…
3abiton on HN
Questions
How to make it useful in practice
From 1 comment
This looks cool but can you explain how to make it useful?
benf76 on HN
Alternatives named
transformer_lens visualizations
From 1 comment
Hey! This is pretty neat, it reminds me of the graphs made by transformer_lens. Cool to see all of these visualization libraries popping up!
JackYoustra on HN
OpenAI and Anthropic sparse autoencoder concept extraction research
From 2 comments
This seems to be what Anthropic and OpenAI did in their research Golden Gate Claude - https://news.ycombinator.com/item?id=40459543 - (60 comments, 16 days ago) Extracting…
SushiHippie on HNThat OpenAI work is more elaborate. It trains an additional network in such a way that it encodes what GPT is doing in terms of activations, but in a more interpretable way…
dimatura on HN