File size: 9,403 Bytes
43d29b8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c5771b4
2a58472
8c42697
c5771b4
2a58472
8c42697
 
 
 
 
 
2a58472
8c42697
2a58472
c5771b4
 
 
 
2a58472
c5771b4
2a58472
8c42697
c5771b4
2a58472
8c42697
 
 
 
2a58472
 
43d29b8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
[
  "tests/test_claim1_needs.py::test_appendix_category_aliases_recorded",
  "tests/test_claim1_needs.py::test_apply_deltas_clamps_and_ignores_unknown_dimensions",
  "tests/test_claim1_needs.py::test_bounded_sum_is_nonnegative_and_clipped",
  "tests/test_claim1_needs.py::test_category_index_is_consistent",
  "tests/test_claim1_needs.py::test_category_means_average_over_subdimensions",
  "tests/test_claim1_needs.py::test_claim1_concordia_scope",
  "tests/test_claim1_needs.py::test_current_values_initialize_across_full_likert_range",
  "tests/test_claim1_needs.py::test_degree_adverb_mapping_matches_appendix_e1",
  "tests/test_claim1_needs.py::test_describe_mentions_every_dimension_and_category",
  "tests/test_claim1_needs.py::test_every_subdimension_has_a_table9_trait",
  "tests/test_claim1_needs.py::test_expected_values_initialize_in_paper_range",
  "tests/test_claim1_needs.py::test_five_major_categories",
  "tests/test_claim1_needs.py::test_four_svo_profiles",
  "tests/test_claim1_needs.py::test_gap_clips_at_zero_when_oversatisfied",
  "tests/test_claim1_needs.py::test_gap_formula_basic",
  "tests/test_claim1_needs.py::test_gap_never_exceeds_s_max",
  "tests/test_claim1_needs.py::test_gaps_cover_all_thirteen_dimensions",
  "tests/test_claim1_needs.py::test_maslow_and_perma_constructs_present",
  "tests/test_claim1_needs.py::test_profile_init_pins_initial_conditions",
  "tests/test_claim1_needs.py::test_qualitative_description_is_monotone",
  "tests/test_claim1_needs.py::test_subdimension_counts_per_category",
  "tests/test_claim1_needs.py::test_svo_prompt_condition_carries_both_halves",
  "tests/test_claim1_needs.py::test_svo_rejects_unknown_orientation",
  "tests/test_claim1_needs.py::test_svo_target_is_stable_while_effective_orientation_moves",
  "tests/test_claim1_needs.py::test_theta_balanced_when_both_signals_vanish",
  "tests/test_claim1_needs.py::test_theta_bounded_to_zero_pi_over_two",
  "tests/test_claim1_needs.py::test_theta_direction_self_vs_other",
  "tests/test_claim1_needs.py::test_thirteen_subdimensions_total",
  "tests/test_engine_and_measure.py::test_action_contract_is_shared_across_architectures",
  "tests/test_engine_and_measure.py::test_all_five_architectures_registered",
  "tests/test_engine_and_measure.py::test_behaviour_text_excludes_internal_state",
  "tests/test_engine_and_measure.py::test_bullying_control_arm_has_no_teacher",
  "tests/test_engine_and_measure.py::test_bullying_intervention_arms_differ_only_by_teacher_goal",
  "tests/test_engine_and_measure.py::test_d2a_has_a_genuine_desire_system",
  "tests/test_engine_and_measure.py::test_each_architecture_produces_an_action[BabyAGI]",
  "tests/test_engine_and_measure.py::test_each_architecture_produces_an_action[D2A]",
  "tests/test_engine_and_measure.py::test_each_architecture_produces_an_action[EduMirror]",
  "tests/test_engine_and_measure.py::test_each_architecture_produces_an_action[LLMob]",
  "tests/test_engine_and_measure.py::test_each_architecture_produces_an_action[ReAct]",
  "tests/test_engine_and_measure.py::test_edumirror_need_update_clamps_and_traces",
  "tests/test_engine_and_measure.py::test_edumirror_selects_argmax_candidate",
  "tests/test_engine_and_measure.py::test_edumirror_survives_malformed_candidate_generation",
  "tests/test_engine_and_measure.py::test_edumirror_uses_three_calls_per_step_and_baselines_fewer",
  "tests/test_engine_and_measure.py::test_election_assigns_all_four_svo_profiles",
  "tests/test_engine_and_measure.py::test_election_interventions_are_environment_level",
  "tests/test_engine_and_measure.py::test_episode_is_reproducible_from_seed",
  "tests/test_engine_and_measure.py::test_every_architecture_gets_identical_persona_text",
  "tests/test_engine_and_measure.py::test_extract_json_handles_real_model_output_shapes[<think>I should return JSON</think>{\"a\": 1}-expected4]",
  "tests/test_engine_and_measure.py::test_extract_json_handles_real_model_output_shapes[Here is the JSON:\\n```\\n{\"a\": 1}\\n```-expected2]",
  "tests/test_engine_and_measure.py::test_extract_json_handles_real_model_output_shapes[Sure! {\"a\": 1} Hope that helps.-expected3]",
  "tests/test_engine_and_measure.py::test_extract_json_handles_real_model_output_shapes[[1, 2, 3]-expected5]",
  "tests/test_engine_and_measure.py::test_extract_json_handles_real_model_output_shapes[```json\\n{\"a\": 1}\\n```-expected1]",
  "tests/test_engine_and_measure.py::test_extract_json_handles_real_model_output_shapes[{\"a\": \"brace } inside string\"}-expected6]",
  "tests/test_engine_and_measure.py::test_extract_json_handles_real_model_output_shapes[{\"a\": 1}-expected0]",
  "tests/test_engine_and_measure.py::test_extract_json_returns_default_on_garbage",
  "tests/test_engine_and_measure.py::test_kindergarten_rejects_impossible_sizes",
  "tests/test_engine_and_measure.py::test_kindergarten_scales_to_paper_group_sizes[15]",
  "tests/test_engine_and_measure.py::test_kindergarten_scales_to_paper_group_sizes[30]",
  "tests/test_engine_and_measure.py::test_kindergarten_scales_to_paper_group_sizes[5]",
  "tests/test_engine_and_measure.py::test_pairwise_compare_randomizes_order_and_maps_back",
  "tests/test_engine_and_measure.py::test_pairwise_compare_returns_none_on_unparseable_verdict",
  "tests/test_engine_and_measure.py::test_rate_behaviour_clamps_out_of_range_scores",
  "tests/test_engine_and_measure.py::test_rate_behaviour_omits_unparseable_metrics_rather_than_defaulting",
  "tests/test_engine_and_measure.py::test_rating_average_matches_paper_aggregation",
  "tests/test_engine_and_measure.py::test_rating_metrics_are_the_four_from_table_1",
  "tests/test_engine_and_measure.py::test_representative_scenarios_cover_paper_themes",
  "tests/test_engine_and_measure.py::test_rses_has_ten_items",
  "tests/test_engine_and_measure.py::test_rses_rejects_wrong_item_count",
  "tests/test_engine_and_measure.py::test_rses_reverse_coding_follows_standard_not_paper_typo",
  "tests/test_engine_and_measure.py::test_rses_score_range",
  "tests/test_engine_and_measure.py::test_run_episode_end_to_end_produces_full_transcript",
  "tests/test_engine_and_measure.py::test_survey_rses_marks_malformed_responses_invalid",
  "tests/test_engine_and_measure.py::test_survey_rses_never_sees_the_transcript",
  "tests/test_engine_and_measure.py::test_svo_ablation_removes_social_value_system",
  "tests/test_engine_and_measure.py::test_svo_slider_rejects_wrong_length",
  "tests/test_engine_and_measure.py::test_svo_slider_scoring_recovers_known_orientations",
  "tests/test_engine_and_measure.py::test_win_rate_matches_equation_4",
  "tests/test_judgecheck.py::test_additive_corruption_keeps_every_original_line",
  "tests/test_judgecheck.py::test_blind_judge_is_reported_as_blind_not_unreadable",
  "tests/test_judgecheck.py::test_corruption_suite_targets_distinct_metrics",
  "tests/test_judgecheck.py::test_destructive_corruption_is_flagged_and_excluded_from_the_verdict",
  "tests/test_judgecheck.py::test_extract_json_handles_reasoning_formats[<think>I should score this 4. {not json}</think>{\"a\": 1}-expected3]",
  "tests/test_judgecheck.py::test_extract_json_handles_reasoning_formats[<think>reasoning that never closes</think>\\n{\"a\": 1}-expected4]",
  "tests/test_judgecheck.py::test_extract_json_handles_reasoning_formats[<|channel|>analysis<|message|>The user wants JSON. Let me think about {braces} and \"quotes\" carefully.<|end|><|start|>assistant<|channel|>final<|message|>{\"Naturalness\": 4}<|return|>-expected0]",
  "tests/test_judgecheck.py::test_extract_json_handles_reasoning_formats[<|channel|>analysis<|message|>thinking...<|end|>{\"a\": 1}-expected2]",
  "tests/test_judgecheck.py::test_extract_json_handles_reasoning_formats[<|channel|>final<|message|>{\"a\": 1}-expected1]",
  "tests/test_judgecheck.py::test_extract_json_handles_reasoning_formats[{\"a\": 1}-expected5]",
  "tests/test_judgecheck.py::test_flatten_all_leaves_no_natural_prose_standing",
  "tests/test_judgecheck.py::test_pairwise_all_unparseable_is_unreadable_not_blind",
  "tests/test_judgecheck.py::test_permutation_corruptions_preserve_every_character",
  "tests/test_judgecheck.py::test_probe_absolute_passes_a_discriminating_judge",
  "tests/test_judgecheck.py::test_probe_absolute_reports_blind_judge_as_blind",
  "tests/test_judgecheck.py::test_probe_pairwise_passes_a_judge_that_prefers_real",
  "tests/test_judgecheck.py::test_probe_pairwise_requires_above_chance_preference",
  "tests/test_judgecheck.py::test_probe_reports_per_metric_target_margins",
  "tests/test_judgecheck.py::test_probe_verdict_requires_every_corruption_to_be_caught",
  "tests/test_judgecheck.py::test_probe_verdict_requires_every_monotone_corruption_to_be_caught",
  "tests/test_judgecheck.py::test_rating_records_raw_only_on_failure",
  "tests/test_judgecheck.py::test_robotic_replaces_actions_but_leaves_narration",
  "tests/test_judgecheck.py::test_role_swap_reattributes_actions_without_inventing_text",
  "tests/test_judgecheck.py::test_shuffle_preserves_content_but_destroys_order",
  "tests/test_judgecheck.py::test_strip_reasoning_prefers_final_channel_over_analysis",
  "tests/test_judgecheck.py::test_unreadable_judge_is_not_reported_as_blind",
  "tests/test_judgecheck.py::test_usage_counts_truncated_responses",
  "tests/test_judgecheck.py::test_valid_judge_gets_valid_verdict",
  "tests/test_judgecheck.py::test_verdict_ignores_a_destructive_corruption_failure"
]