It's great that you're aiming for clean code architecture and considering localization at a feature level.
While clean code, it's important not to over-complicate or tightly couple solutions. The common structure of having a centralized localization file is already optimized and sticks to clean code principles. Before restructuring, ask yourself:

• If the same string appears across multiple features, will you duplicate it in each file?
• When scaling in the future, how much extra effort will be needed to search through various files to update translations?

From my experience, a centralized approach is easier because many strings are reused. It simplifies updates and makes localization management more efficient by handling everything from a single source.

Example: