| [ |
| "tests/test_capture.py::test_load_entrypoint_registers_module_for_hydra_relative_configs", |
| "tests/test_capture.py::test_run_main_points_hydra_at_sibling_config_directory", |
| "tests/test_capture.py::test_teacher_capture_accepts_upstream_python_list_samples", |
| "tests/test_conditional_evaluation.py::test_conditional_ppl_excludes_prompt_tokens_from_nll", |
| "tests/test_conditional_evaluation.py::test_conditional_texts_decode_only_evaluation_suffix_for_long_canvas", |
| "tests/test_entropy.py::test_entropy_in_nats", |
| "tests/test_entropy.py::test_mean_entropy_is_per_sample_and_removes_only_documented_padding", |
| "tests/test_entropy.py::test_mean_entropy_rejects_a_row_containing_only_excluded_ids", |
| "tests/test_entropy.py::test_metric_cli_atomically_labels_partial_fixture_result", |
| "tests/test_entropy.py::test_metric_cli_requires_explicit_partial_override", |
| "tests/test_entropy.py::test_partial_custom_exclusion_is_labelled_nonbaseline", |
| "tests/test_entropy.py::test_production_cli_rejects_custom_special_token_exclusions", |
| "tests/test_perplexity.py::test_aggregate_nll_rejects_empty_or_nonfinite_aggregates[parts0]", |
| "tests/test_perplexity.py::test_aggregate_nll_rejects_empty_or_nonfinite_aggregates[parts1]", |
| "tests/test_perplexity.py::test_aggregate_nll_rejects_empty_or_nonfinite_aggregates[parts2]", |
| "tests/test_perplexity.py::test_aggregate_nll_rejects_empty_or_nonfinite_aggregates[parts3]", |
| "tests/test_perplexity.py::test_compute_gen_ppl_is_batch_invariant_and_scores_final_short_batch", |
| "tests/test_perplexity.py::test_compute_gen_ppl_rejects_any_sample_without_a_next_token_target", |
| "tests/test_perplexity.py::test_compute_gen_ppl_rejects_context_beyond_gpt2_limit", |
| "tests/test_perplexity.py::test_compute_gen_ppl_rejects_corpus_without_next_token_targets", |
| "tests/test_perplexity.py::test_compute_gen_ppl_rejects_non_right_padded_attention_mask", |
| "tests/test_perplexity.py::test_compute_gen_ppl_rejects_scorer_token_count_drift", |
| "tests/test_perplexity.py::test_loader_forces_local_snapshots_and_offline_mode", |
| "tests/test_perplexity.py::test_ppl_uses_token_weighted_nll", |
| "tests/test_perplexity.py::test_resolver_binds_model_and_tokenizer_to_pinned_local_snapshots", |
| "tests/test_perplexity.py::test_resolver_rejects_download_record_at_another_revision", |
| "tests/test_perplexity.py::test_resolver_rejects_non_lowercase_hex_revision", |
| "tests/test_perplexity.py::test_resolver_rejects_symlinked_snapshot_ancestor", |
| "tests/test_perplexity.py::test_resolver_rejects_unknown_download_manifest_schema", |
| "tests/test_perplexity.py::test_torch_scorer_shifts_masks_and_sums_causal_nll", |
| "tests/test_self_bleu.py::test_equidistant_reference_length_tie_chooses_shorter", |
| "tests/test_self_bleu.py::test_leave_one_out_clipping_uses_second_max_for_unique_top_owner", |
| "tests/test_self_bleu.py::test_method1_keeps_zero_when_candidate_has_no_unigram_reference_match", |
| "tests/test_self_bleu.py::test_self_bleu_builds_each_row_order_counter_once", |
| "tests/test_self_bleu.py::test_self_bleu_detects_mode_collapse", |
| "tests/test_self_bleu.py::test_self_bleu_has_explicit_deterministic_standard_conventions", |
| "tests/test_self_bleu.py::test_self_bleu_rejects_missing_reference_or_empty_tokens[texts0]", |
| "tests/test_self_bleu.py::test_self_bleu_rejects_missing_reference_or_empty_tokens[texts1]", |
| "tests/test_self_bleu.py::test_self_bleu_rejects_missing_reference_or_empty_tokens[texts2]", |
| "tests/test_training_recipes.py::test_di4c_hf_small_student_init_preserves_configured_teacher" |
| ] |
