Feature: Get the value for rotary base from the hugging face config, only for Qwen for now.

[Gusanidas]

Description

Two line change so that the rope_base parameter in the configuration is taken from the hugging face config if available.

When loading a HookedTransformer from a huggin face model in one of the known architectures, it may be the case that the rope base (also named rope theta) has a different value for the hugging_face model that the default for that architecture.

This is the case for example when loading "deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B" into the "Qwen 2.5" architecture.

The code already took the values for "d_vocab" and "load_in_4bit" from the hf config, I am extending it also to rope base.

I have limited it to Qwen models because otherwise many different test fail, but this is a change that will be usefull and accepted I can work on doing the general case, and adding the relevant tests, etc

Fixes # (issue)

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)
• [x ] New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

Screenshots

Please attach before and after screenshots of the change if applicable.

Checklist:

• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
• I have not rewritten tests relating to key interfaces which would affect backward compatibility