Aurélien Dupays Dexemple 💾
@aureliendupaysdexemple.com
Product Engineer @liveblocks.io 🚀
Reposted by Aurélien Dupays Dexemple 💾
First you need to understand embeddings.
Embeddings turns words into lists of numbers, a vector, that represent how closely related they are, semantically.
Spelling doesn't matter, for example "cat" and "dog" are closely linked, but "cat" and "car" are not.
Embeddings turns words into lists of numbers, a vector, that represent how closely related they are, semantically.
Spelling doesn't matter, for example "cat" and "dog" are closely linked, but "cat" and "car" are not.
![// ✅ Closely related "cat": [0.25, 0.84, ...] "dog": [0.28, 0.81, ...] // ❌ Not related "cat": [0.25, 0.84, ...] "car": [0.91, 0.12, ...]](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:i3yxxcbu3dzxj7dpb2mfriag/bafkreif5ffj3u6un5ckkh24spv2lp37dnozkhhyuy4l7gjyvf5lgqi5w3i@jpeg)
August 7, 2025 at 2:56 PM
First you need to understand embeddings.
Embeddings turns words into lists of numbers, a vector, that represent how closely related they are, semantically.
Spelling doesn't matter, for example "cat" and "dog" are closely linked, but "cat" and "car" are not.
Embeddings turns words into lists of numbers, a vector, that represent how closely related they are, semantically.
Spelling doesn't matter, for example "cat" and "dog" are closely linked, but "cat" and "car" are not.
Let’s go 🚀🚀
June 18, 2025 at 7:28 AM
Let’s go 🚀🚀