Critique Fine-Tuning: Learning to Critique is More Effective than Learning to Imitate
@css_nlp
Please open Telegram to view this post
VIEW IN TELEGRAM
👍3❤1🆒1
Diagram Generator
A tool for generating diagrams using LLM, with a Python backend API and React frontend.
This project was purely an experiment, it is entirely generated with guidance from myself.
Otherwise known as "vibe coding".
Would I recommend it? No, in fact don't even credit me if you use this code.
I do however believe that as engineers we need to explore options we disagree with in order to widen our perspective. I did in fact learn a lot about how LLMs "think" during this project and am glad to have done it regardless of how frustrating it was.
@css_nlp
Please open Telegram to view this post
VIEW IN TELEGRAM
GitHub
GitHub - minimalefforttech/diagram_generator: LLM Powered Diagram Generator
LLM Powered Diagram Generator. Contribute to minimalefforttech/diagram_generator development by creating an account on GitHub.
👍5
Please open Telegram to view this post
VIEW IN TELEGRAM
Linkedin
Stanford Top 2% list: Misleading and Distorting Science | Achal Agrawal, PhD posted on the topic | LinkedIn
The 'Stanford Top 2%' list has released and the brag-a-thon has started. Even when the list explicitly has a column for retractions, being in the list is somehow equated to quality.
We see that the percentage of authors in the list with at least 1 retraction…
We see that the percentage of authors in the list with at least 1 retraction…
🤣2
Large Language Model Hacking:Quantifying the Hidden Risks of Using LLMs for Text Annotation
@css_nlp
Please open Telegram to view this post
VIEW IN TELEGRAM