Hacker News new | past | comments | ask | show | jobs | submit login

A vector representation of a text based on its meaning, generated by an ML model. It's the data format used as input for LLMs and can be used to compare the meaning of texts by comparing the vectors with each other.



Is it as simple as the euclidean distances of the vectors in N-dimensional space is intended to approximate the difference in meaning between two texts?


Yes, the (squared) euclidean distance is one of the three common methods, as well as cosine similarity and dot product.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: