[Feature] Update embedding model to text-embedding-3-large

[athrael-soju]

Is this your first time submitting a feature request?

• I have searched the existing issues, and I could not find an existing issue for this feature
• I am requesting a straightforward extension of existing functionality

Describe the feature

This is a very minor request to update the embedding model used to text-embedding-3-large here

As it's the most capable 3rd generation embedding model, it would be nice to see how it performs over text-embedding-ada-002

Note:

• This would require an Output Dimension of 3,072, so the index would have to be recreated as such.
• Also, text-embedding-3-large is just slightly more expensive, at $0.13 / 1M tokens, but nothing to write home about.

Describe alternatives you've considered

text-embedding-3-small can also be an option, as it's still has Increased performance over the 2nd generation ada embedding model and retains the vector dimension at 1536, so it would be a matter of just replacing the existing model in the code.

Note:

• text-embedding-3-small is also much cheaper to use, at $0.02 / 1M tokens

Who will this benefit?

Embedding V3 large is the most capable embedding model for both english and non-english tasks, with expected precision increase in semantic search tasks.

Are you interested in contributing this feature?

Yes

Anything else?

Happy to make this change, if needed.