Cursor for learning.
An AI-native document reader that answers questions with page citations, turns highlights into flashcards, and connects you to two million arXiv papers.
No credit card required · Works with any PDF or EPUB
3.2 Attention
An attention function can be described as mapping a query and a set of key-value pairs to an output, where the query, keys, values, and output are all vectors.
The output is computed as a weighted sum of the values, where the weight assigned to each value is computed by a compatibility function of the query with the corresponding key.
We call our particular attention "Scaled Dot-Product Attention". The input consists of queries and keys of dimension d_k, and values of dimension d_v.
RorobotRorobotBuilt for people who read to learn
Ask anything. Get cited answers.
No more re-reading to find that one passage. Ask a question in plain English, get an answer grounded in your document—with the exact page number.
Try it freeRorobotTurn reading into remembering
One highlight becomes a flashcard. One paper becomes a quiz. Your reading stops being a one-time event and becomes lasting knowledge.
Try it freeTwo million papers, one click away
Search arXiv without leaving the app. Find a paper, open it, and start reading with full AI tools—no downloading, no tab-switching, no friction.
Try it freeHow it works
Three steps to deeper reading
Drop your document
Upload a PDF, EPUB, or open any paper from arXiv. AI indexes it in the background while you start reading.
Read with superpowers
Highlight passages, ask questions about any section, and get answers that cite exact page numbers.
Actually retain it
Your highlights become flashcards. Your papers become quizzes. Everything lives in one searchable knowledge hub.