Part 3 of Decoding LLMs, a series on what happens after a model produces its raw output. The first knob becomes the only knob When the output is wrong, is reaching for the temperature dial the...

Part 2 of Decoding LLMs, a series on what happens after a model produces its raw output. A model that should repeat itself Same prompt, temperature zero, sent twice — why are the answers diffe...

Part 1 of Decoding LLMs, a series on what happens after a model produces its raw output. Picking a word out of fifty thousand You type a prompt, hit enter, and a word appears. Where does that ...

This post was originally published on Medium on July 7, 2018. View original article. There’s this quote I love from Bill Gates: “I choose a lazy person to do a hard job. Because a lazy person wil...

This post was originally published on Medium on June 29, 2018. View original article. Get ready to classify some data! Logistic regression is a supervised ML algorithm. A bit of calculus, probab...

This post was originally published on Medium on June 21, 2018. View original article. Probability Basics Probability is the chance of an event $A$ occurring out of all possible events in the sam...

This post was originally published on Medium on June 18, 2018. View original article. Introduction to Machine Learning Before I get into gradient descent, let me quickly walk through how I think...