Amelia Wattenberger

Here's a close-up of a fish— we can see the scales, the texture of the fins, the faint reflection in the eye. Zoomed out a bit, the details are faded slightly. The surrounding ocean comes into view. Zooming out more, the fish is just a few splashes of color amongst a school of fish. Maps work in a similar way. Looking at a section of the National Zoo, we can see paths, buildings, and even animal enclosures! Zoomed out a little, and now it’s about neighborhoods—parks, stores, and streets. Keep zooming out and you’re looking at a city, with just the highways and terrain. Each level gives you completely different information, depending on what Google thinks the user might be interested in. Maps are a true masterclass for visualizing the same information in a variety of ways. So far, these are familiar examples. What if we applied the same principles to text? Let's take a look at the opening paragraph of The Metamorphosis. As Gregor Samsa awoke one morning from uneasy dreams he found himse

Fish eyes We often think of the same information at different levels or abstraction. Here's a simple example: Here's a close-up of a fish— we can see the scales, the texture of the fins, the faint reflection in the eye. Zoomed out a bit, the details are faded slightly. The surrounding ocean comes into view. Zooming out more, the fish is just a few splashes of color amongst a school of fish. Maps work in a similar way. Looking at a section of the National Zoo, we can see paths, buildings, and even animal enclosures! Zoomed out a little, and now it’s about neighborhoods—parks, stores, and street

saved by

• Annie Zhou
• Aileen Luo
• Aaron Pham
• Wilbur Zhang
• Sudarsh K
• Flora Guo
• Yuemin Huang
• Tim Bols

related reading

• How can we develop transformative tools for thought?numinous.productions
• How to Use LLMs Better - Learn How To Learnlearnhowtolearn.org
• Prism: mapping interpretable concepts and features in a latent space of language | thesephist.comthesephist.com
• Transformer Explainer: LLM Transformer Model Visually Explainedpoloclub.github.io
• Deep Dive into LLMs like ChatGPTyoutube.com
• LLMs are a tool for thoughtwattenberger.com
• LLM Daydreaming · Gwern.netgwern.net
• Rotating the Space: On LLMs as a Medium for Thought | rotating_the_spacesbgeoaiphd.github.io
• Against LLM Reductionism — LessWronglesswrong.com
• How Does A Blind Model See The Earth? — LessWronglesswrong.com
• Best explanations of how LLMs work | Roman Vorushinvorushin.github.io
• Do large language models understand us? | by Blaise Aguera y Arcas | Mediummedium.com