The Truth of AI retweeted
Improving LLMs' math reasoning ability is actually not as easy as what people anticipated. There are some interesting observations:
1. Using textbooks does not improve models' reasoning a lot. We tried to use the 800 high-quality math textbooks from Mathpile to continue training
The Truth of AI retweeted
Here, in full directly on Twitter, is "A Hackers' Guide to Language Models". This 90 minute tutorial is designed to be the one place I point coders at when they ask "hey, tell me everything I need to know about LLMs!"
It covers both @OpenAI models and open source ones in depth.
The Truth of AI retweeted
"startups have shamelessly re-branded rudimentary machine-learning algorithms as the dawn of the singularity... Welcome to the artificial intelligence bullshit-industrial complex”
I wonder if the commercialisation of #AIethics is amplifying the bullshit,
gen.medium.com/the-bs-indust…
The Truth of AI retweeted
Why we should focus on weak artificial intelligence for the moment bdtechtalks.com/2018/04/23/s…
#ArtificialIntelligence #MachineLearning #AI
The Truth of AI retweeted
An audit of a resume screening algorithm found that the two factors most indicative of job performance were being named Jared and playing high school lacrosse
qz.com/1427621/companies-are…

ALT Mark J. Girouard, an employment attorney at Nilan Johnson Lewis, says one of his clients was vetting a company selling a resume screening tool, but didn’t want to make the decision until they knew what the algorithm was prioritizing in a person’s CV. After an audit of the algorithm, the resume screening company found that the algorithm found two factors to be most indicative of job performance: their name was Jared,
The Truth of AI retweeted
My ML seminar course is focusing on "transfer learning" this term. Now reading usenix.org/system/files/conf… Shows how transfer learning (here, taking common pre-trained deep-learning models & fine-tuning them) results in vulnerability to black-box adversarial attacks. Worth reading.







