Improve model config/tokenizer format defaults

[Daniel-Schenker]

Description

This PR improves upon the faulty logic used in #1027.

The previous fix attempted to exclude Llama models from mistral-format detection by checking whether params.json existed only inside an original/ subdirectory, using two calls to any_pattern_in_repo_files. This did not work as the vLLM implementation of any_pattern_in_repo_files delegates to list_filtered_repo_files, which matches patterns using fnmatch.fnmatch(os.path.basename(file), pattern). This always returns false since os.path.basename is applied to every file path before fnmatch runs. Llama models with only original/params.json were still being assigned mistral format.

fix: correctly detect root-level params.json for mistral format detection.

Replace both any_pattern_in_repo_files calls with a single list_repo_files call followed by an exact string membership test.

Related Issues

Test Plan

Checklist

• I have read the contributing guidelines
• My code follows the project's code style (run bash format.sh)
• I have added tests for my changes (if applicable)
• I have updated the documentation (if applicable)
• My commits include a Signed-off-by: line (DCO compliance)

[github-actions]

👋 Hi! Thank you for contributing.
Just a reminder: Make sure that your code passes all the linting checks, otherwise your PR won't be able to be merged. To do so, run ./format.sh.
Now you are good to go 🚀.

We also recommend installing prek and configuring it to check your code before every local commit.