    Finished `test` profile [unoptimized + debuginfo] target(s) in 0.46s
     Running tests/label_precision_benchmark.rs (target/debug/deps/label_precision_benchmark-c64808f7653fac86)

running 1 test
label_precision_haiku: model=claude-haiku-4-5-20251001
haiku ingest: episode_id=1 entities=16 facts=3
haiku extracted (18): [("person", "Entity"), ("new generator technology", "Entity"), ("Africa", "Entity"), ("Amazon", "Organisation"), ("Amazon Robotics", "Organisation"), ("Boston", "Entity"), ("Boston Consulting Group", "Entity"), ("COVID", "Entity"), ("C-suite", "Entity"), ("France", "Entity"), ("Hi I'm Ria Thank", "Entity"), ("Japanese", "Entity"), ("Morocco", "Entity"), ("Northeastern University While", "Entity"), ("Ria", "Person"), ("Robotics", "Entity"), ("South Korea", "Entity"), ("Tableau", "Entity")]
label_precision (Haiku): 30.0% (3/10 ground-truth pairs matched, 15 placeholder)

thread 'label_precision_haiku_on_mock_interview' (9452906) panicked at crates/kremory/tests/label_precision_benchmark.rs:539:5:
Haiku label_precision 30.0% (3/10) below 65% R1 threshold. 15 entities had non-canonical labels. Model: claude-haiku-4-5-20251001
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
test label_precision_haiku_on_mock_interview ... FAILED

failures:

failures:
    label_precision_haiku_on_mock_interview

test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 5 filtered out; finished in 20.05s

error: test failed, to rerun pass `-p kremory --test label_precision_benchmark`
