fix: adapt to llama.cpp changes

[giladgd]

Description of change

• fix: adapt to llama.cpp changes
• fix: pad the context size to align with the implementation in llama.cpp
• fix: fallback to other available binaries on load failure in internal llama getter without backend
• fix: bugs

You may notice that the context size that gets created is sometimes slightly larger than what you specify (by up to 256) - this happens since llama.cpp pads the size of the context to align to multiples of 256 for performance reasons.
You can check the actual context size that gets created using the .contextSize property on the created context or any of its sequences.

Pull-Request Checklist

• Code is up-to-date with the master branch
• npm run format to apply eslint formatting
• npm run test passes with this change
• This pull request links relevant issues as Fixes #0000
• There are new or updated unit tests validating the change
• Documentation has been updated to reflect this change
• The new commits and pull request title follow conventions explained in pull request guidelines (PRs that do not follow this convention will not be merged)