Understand, visualize, and refine prompts with confidence
Hi, I'm Oli. I started building and writing oliprompts while competing in the 2022 Feedback Prize Kaggle competition, visualizing token probabilities for language models (Longformer, BigBird, etc.). I wrote the guides to learn and understand the newest techniques. Now I'm expanding this with my prompt library which I use mostly for coding. It's a mess now but I'll organize it more soon.
Prompt Visualization
See how techniques like role prompting actually engineer token probability landscapes. Originally built for analyzing Longformer, BigBird, and Roberta architectures, now adapted for understanding modern language models with live 3D probability charts.
Open toolPrompt Library
Templates from real AI workflows I use for building. Try them, refine them, and visualize how changes affect output—you'll see how negative prompting often increases the odds of getting exactly what you're trying to avoid.
View templatesLearning Resources
Prompt engineering can be intuitive, but understanding which techniques are for us versus for the models makes all the difference. These guides explain the "why" behind each approach with practical examples and decision frameworks.
Start learningGet started quickly
New to prompting or refining advanced workflows? Pick a resource and follow the guided steps to iterate with intention.
Featured topics
A few places to dive deeper when you want to level up technique and strategy.