Fix Loewner threshold sharing across matrices in a batch

[lucas-heck]

Problem

get_loewner_threshold used the largest absolute eigenvalue across the entire batch to scale a shared threshold, so large large eigenvalues could affect unrelated matrices during gradient computation (quite rare thought).

Fix

Compute the threshold independently for each matrix using its own largest absolute eigenvalue.

Validation

Added regression test to check that one matrix won't influence another.

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!