Add test assertions for label attention attributions shape and presence

[Copilot]

Addresses review feedback requesting assertions to catch label attention wiring/shape regressions.

Changes

• Added assertions in run_full_pipeline test helper to validate label attention attributions:
  • When enabled: verify attributions exist and have shape (batch_size, n_head, num_classes, seq_len)
  • When disabled: verify attributions are None

# Test label attention assertions
if label_attention_enabled:
    assert predictions["label_attention_attributions"] is not None
    label_attention_attributions = predictions["label_attention_attributions"]
    expected_shape = (
        len(sample_text_data),  # batch_size
        model_params["n_head"],  # n_head
        model_params["num_classes"],  # num_classes
        tokenizer.output_dim,  # seq_len
    )
    assert label_attention_attributions.shape == expected_shape
else:
    assert predictions.get("label_attention_attributions") is None

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.