gooo-joint-path-tiny-v1 / study-report.json
asketeddy's picture
Publish six own random-initialized Gooo joint/independent models and audited SDK intermediate results
44bf2bc verified
Raw History Blame Contribute Delete
222 kB
{
"actual_model_predictions_all_stages": 25471,
"benchmark_predictions_including_allocation_probe": 12012,
"calibration": {
"independent-fp32": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2925,
"initial_complete_functions": 154,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 400,
"actual_model_predictions": 1380,
"actual_feedback_predictions": 612,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 752,
"bilingual_initial_mask_disagreements": 180,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
154,
112,
66,
52
],
"complete_function_curve_budgets_1_to_4": [
154,
266,
332,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2925,
4484,
5423,
6144
],
"summed_observed_prediction_ns": 13253190,
"summed_initial_probability_mass_outside_full_target": 263.40002416328724,
"summed_initial_full_target_nll": 525.1140375810159
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 283,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 28,
"actual_model_predictions": 107,
"actual_feedback_predictions": 43,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 68,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
8,
4,
4
],
"complete_function_curve_budgets_1_to_4": [
16,
24,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
283,
398,
454,
512
],
"summed_observed_prediction_ns": 1025626,
"summed_initial_probability_mass_outside_full_target": 19.10443247631663,
"summed_initial_full_target_nll": 43.2229110057714
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 283,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 16,
"actual_model_predictions": 96,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 80,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
16,
0,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
32,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
283,
512,
512,
512
],
"summed_observed_prediction_ns": 863960,
"summed_initial_probability_mass_outside_full_target": 19.687865571513182,
"summed_initial_full_target_nll": 45.46758740092145
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 294,
"initial_complete_functions": 18,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 27,
"actual_model_predictions": 101,
"actual_feedback_predictions": 37,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 69,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
18,
6,
3,
5
],
"complete_function_curve_budgets_1_to_4": [
18,
24,
27,
32
],
"best_passing_case_curve_budgets_1_to_4": [
294,
390,
435,
512
],
"summed_observed_prediction_ns": 993953,
"summed_initial_probability_mass_outside_full_target": 22.950683253568936,
"summed_initial_full_target_nll": 40.61678484472798
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 195,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 29,
"actual_model_predictions": 122,
"actual_feedback_predictions": 58,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 67,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
11,
9,
0
],
"complete_function_curve_budgets_1_to_4": [
12,
23,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
195,
371,
512,
512
],
"summed_observed_prediction_ns": 1162706,
"summed_initial_probability_mass_outside_full_target": 23.849738121665013,
"summed_initial_full_target_nll": 44.54739746654518
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 148,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 49,
"actual_model_predictions": 132,
"actual_feedback_predictions": 68,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 47,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
7,
9,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
15,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
148,
265,
398,
512
],
"summed_observed_prediction_ns": 1238371,
"summed_initial_probability_mass_outside_full_target": 23.59054887476337,
"summed_initial_full_target_nll": 43.87110768353832
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 156,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 30,
"actual_model_predictions": 122,
"actual_feedback_predictions": 58,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 66,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
18,
6,
0
],
"complete_function_curve_budgets_1_to_4": [
8,
26,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
156,
420,
512,
512
],
"summed_observed_prediction_ns": 1119791,
"summed_initial_probability_mass_outside_full_target": 23.618873349610595,
"summed_initial_full_target_nll": 43.69315234621296
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 353,
"initial_complete_functions": 14,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 31,
"actual_model_predictions": 109,
"actual_feedback_predictions": 45,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 65,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
14,
9,
5,
4
],
"complete_function_curve_budgets_1_to_4": [
14,
23,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
353,
431,
478,
512
],
"summed_observed_prediction_ns": 1102625,
"summed_initial_probability_mass_outside_full_target": 17.488844136684623,
"summed_initial_full_target_nll": 44.11157948999173
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 438,
"initial_complete_functions": 18,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 20,
"actual_model_predictions": 102,
"actual_feedback_predictions": 38,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 76,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
18,
9,
4,
1
],
"complete_function_curve_budgets_1_to_4": [
18,
27,
31,
32
],
"best_passing_case_curve_budgets_1_to_4": [
438,
467,
501,
512
],
"summed_observed_prediction_ns": 1003335,
"summed_initial_probability_mass_outside_full_target": 17.437517033614686,
"summed_initial_full_target_nll": 43.31409860785471
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 170,
"initial_complete_functions": 10,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 44,
"actual_model_predictions": 124,
"actual_feedback_predictions": 60,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 52,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
10,
8,
6,
8
],
"complete_function_curve_budgets_1_to_4": [
10,
18,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
170,
313,
399,
512
],
"summed_observed_prediction_ns": 1234497,
"summed_initial_probability_mass_outside_full_target": 23.961364568377736,
"summed_initial_full_target_nll": 44.23390374255849
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 248,
"initial_complete_functions": 14,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 39,
"actual_model_predictions": 116,
"actual_feedback_predictions": 52,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 57,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
14,
4,
7,
7
],
"complete_function_curve_budgets_1_to_4": [
14,
18,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
248,
303,
409,
512
],
"summed_observed_prediction_ns": 1096376,
"summed_initial_probability_mass_outside_full_target": 23.898001494184285,
"summed_initial_full_target_nll": 43.978782570753594
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 145,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 131,
"actual_feedback_predictions": 67,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
145,
273,
398,
512
],
"summed_observed_prediction_ns": 1235623,
"summed_initial_probability_mass_outside_full_target": 23.993335421515553,
"summed_initial_full_target_nll": 44.35708308656053
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 212,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 39,
"actual_model_predictions": 118,
"actual_feedback_predictions": 54,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 57,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
8,
5,
7
],
"complete_function_curve_budgets_1_to_4": [
12,
20,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
212,
341,
415,
512
],
"summed_observed_prediction_ns": 1176327,
"summed_initial_probability_mass_outside_full_target": 23.818819861472818,
"summed_initial_full_target_nll": 43.69964933557976
}
}
},
"independent-ptq_ternary": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2755,
"initial_complete_functions": 144,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 438,
"actual_model_predictions": 1418,
"actual_feedback_predictions": 650,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 714,
"bilingual_initial_mask_disagreements": 168,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
144,
95,
92,
53
],
"complete_function_curve_budgets_1_to_4": [
144,
239,
331,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2755,
4159,
5402,
6144
],
"summed_observed_prediction_ns": 15562753,
"summed_initial_probability_mass_outside_full_target": 264.72765764348514,
"summed_initial_full_target_nll": 525.8331452424557
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 283,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 24,
"actual_model_predictions": 104,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 72,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
8,
8,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
24,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
283,
399,
512,
512
],
"summed_observed_prediction_ns": 1078792,
"summed_initial_probability_mass_outside_full_target": 19.27240256645411,
"summed_initial_full_target_nll": 43.70968475713822
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 283,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 23,
"actual_model_predictions": 110,
"actual_feedback_predictions": 46,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 73,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
9,
7,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
25,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
283,
413,
512,
512
],
"summed_observed_prediction_ns": 1181168,
"summed_initial_probability_mass_outside_full_target": 19.90529919010544,
"summed_initial_full_target_nll": 44.27159617898773
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 133,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 44,
"actual_model_predictions": 124,
"actual_feedback_predictions": 60,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 52,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
12,
4,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
20,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
133,
327,
389,
512
],
"summed_observed_prediction_ns": 1554212,
"summed_initial_probability_mass_outside_full_target": 23.235460226126936,
"summed_initial_full_target_nll": 41.68152078748247
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 196,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 120,
"actual_feedback_predictions": 56,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
8,
12,
0
],
"complete_function_curve_budgets_1_to_4": [
12,
20,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
196,
326,
512,
512
],
"summed_observed_prediction_ns": 1280623,
"summed_initial_probability_mass_outside_full_target": 23.941642513172376,
"summed_initial_full_target_nll": 44.344627617161045
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 148,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 128,
"actual_feedback_predictions": 64,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
148,
274,
397,
512
],
"summed_observed_prediction_ns": 1351458,
"summed_initial_probability_mass_outside_full_target": 23.650489271657296,
"summed_initial_full_target_nll": 44.06002748528022
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 214,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 37,
"actual_model_predictions": 122,
"actual_feedback_predictions": 58,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 59,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
7,
9,
4
],
"complete_function_curve_budgets_1_to_4": [
12,
19,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
214,
320,
454,
512
],
"summed_observed_prediction_ns": 1312297,
"summed_initial_probability_mass_outside_full_target": 23.76501813433186,
"summed_initial_full_target_nll": 43.53052557066708
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 353,
"initial_complete_functions": 14,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 36,
"actual_model_predictions": 114,
"actual_feedback_predictions": 50,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 60,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
14,
4,
10,
4
],
"complete_function_curve_budgets_1_to_4": [
14,
18,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
353,
438,
478,
512
],
"summed_observed_prediction_ns": 1296044,
"summed_initial_probability_mass_outside_full_target": 17.572842196378826,
"summed_initial_full_target_nll": 44.26594206199383
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 438,
"initial_complete_functions": 18,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 18,
"actual_model_predictions": 100,
"actual_feedback_predictions": 36,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 78,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
18,
10,
4,
0
],
"complete_function_curve_budgets_1_to_4": [
18,
28,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
438,
478,
512,
512
],
"summed_observed_prediction_ns": 1063375,
"summed_initial_probability_mass_outside_full_target": 17.55521231734092,
"summed_initial_full_target_nll": 42.98990900789376
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 140,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 47,
"actual_model_predictions": 128,
"actual_feedback_predictions": 64,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 49,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
9,
7
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
140,
283,
411,
512
],
"summed_observed_prediction_ns": 1386664,
"summed_initial_probability_mass_outside_full_target": 23.977058211833864,
"summed_initial_full_target_nll": 44.33217621037191
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 205,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 42,
"actual_model_predictions": 120,
"actual_feedback_predictions": 56,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 54,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
6,
6,
8
],
"complete_function_curve_budgets_1_to_4": [
12,
18,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
205,
309,
396,
512
],
"summed_observed_prediction_ns": 1299754,
"summed_initial_probability_mass_outside_full_target": 23.94919347170267,
"summed_initial_full_target_nll": 44.1872961286787
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 145,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 130,
"actual_feedback_predictions": 66,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
145,
272,
400,
512
],
"summed_observed_prediction_ns": 1414032,
"summed_initial_probability_mass_outside_full_target": 23.983555643931894,
"summed_initial_full_target_nll": 44.3405321750111
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 217,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 39,
"actual_model_predictions": 118,
"actual_feedback_predictions": 54,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 57,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
7,
7,
6
],
"complete_function_curve_budgets_1_to_4": [
12,
19,
26,
32
],
"best_passing_case_curve_budgets_1_to_4": [
217,
320,
429,
512
],
"summed_observed_prediction_ns": 1344334,
"summed_initial_probability_mass_outside_full_target": 23.919483900448835,
"summed_initial_full_target_nll": 44.11930726178957
}
}
},
"independent-qat_ternary": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2833,
"initial_complete_functions": 148,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 415,
"actual_model_predictions": 1397,
"actual_feedback_predictions": 629,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 737,
"bilingual_initial_mask_disagreements": 166,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
148,
114,
65,
57
],
"complete_function_curve_budgets_1_to_4": [
148,
262,
327,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2833,
4382,
5337,
6144
],
"summed_observed_prediction_ns": 15247841,
"summed_initial_probability_mass_outside_full_target": 263.43281332204407,
"summed_initial_full_target_nll": 522.2127398246924
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 283,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 31,
"actual_model_predictions": 110,
"actual_feedback_predictions": 46,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 65,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
7,
3,
6
],
"complete_function_curve_budgets_1_to_4": [
16,
23,
26,
32
],
"best_passing_case_curve_budgets_1_to_4": [
283,
382,
424,
512
],
"summed_observed_prediction_ns": 1235583,
"summed_initial_probability_mass_outside_full_target": 19.12043977919033,
"summed_initial_full_target_nll": 42.83170988137251
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 283,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 16,
"actual_model_predictions": 96,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 80,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
16,
0,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
32,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
283,
512,
512,
512
],
"summed_observed_prediction_ns": 1048411,
"summed_initial_probability_mass_outside_full_target": 19.753837449680002,
"summed_initial_full_target_nll": 44.932851107514765
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 198,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 36,
"actual_model_predictions": 115,
"actual_feedback_predictions": 51,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 60,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
10,
4,
6
],
"complete_function_curve_budgets_1_to_4": [
12,
22,
26,
32
],
"best_passing_case_curve_budgets_1_to_4": [
198,
357,
420,
512
],
"summed_observed_prediction_ns": 1239832,
"summed_initial_probability_mass_outside_full_target": 22.90731458391581,
"summed_initial_full_target_nll": 40.47629878516883
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 257,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 25,
"actual_model_predictions": 114,
"actual_feedback_predictions": 50,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 71,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
7,
9,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
23,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
257,
371,
512,
512
],
"summed_observed_prediction_ns": 1288706,
"summed_initial_probability_mass_outside_full_target": 23.77594026695118,
"summed_initial_full_target_nll": 44.11102196754362
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 148,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 52,
"actual_model_predictions": 133,
"actual_feedback_predictions": 69,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 44,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
7,
6,
11
],
"complete_function_curve_budgets_1_to_4": [
8,
15,
21,
32
],
"best_passing_case_curve_budgets_1_to_4": [
148,
266,
356,
512
],
"summed_observed_prediction_ns": 1420457,
"summed_initial_probability_mass_outside_full_target": 23.637616891284992,
"summed_initial_full_target_nll": 43.59599042773189
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 214,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 28,
"actual_model_predictions": 114,
"actual_feedback_predictions": 50,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 68,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
14,
4,
2
],
"complete_function_curve_budgets_1_to_4": [
12,
26,
30,
32
],
"best_passing_case_curve_budgets_1_to_4": [
214,
422,
481,
512
],
"summed_observed_prediction_ns": 1216253,
"summed_initial_probability_mass_outside_full_target": 23.642319032168224,
"summed_initial_full_target_nll": 43.46101273353078
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 352,
"initial_complete_functions": 14,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 26,
"actual_model_predictions": 107,
"actual_feedback_predictions": 43,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 70,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
14,
12,
4,
2
],
"complete_function_curve_budgets_1_to_4": [
14,
26,
30,
32
],
"best_passing_case_curve_budgets_1_to_4": [
352,
435,
493,
512
],
"summed_observed_prediction_ns": 1149702,
"summed_initial_probability_mass_outside_full_target": 17.546220340549805,
"summed_initial_full_target_nll": 43.79877485824531
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 438,
"initial_complete_functions": 18,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 19,
"actual_model_predictions": 100,
"actual_feedback_predictions": 36,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 77,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
18,
10,
3,
1
],
"complete_function_curve_budgets_1_to_4": [
18,
28,
31,
32
],
"best_passing_case_curve_budgets_1_to_4": [
438,
478,
506,
512
],
"summed_observed_prediction_ns": 1062255,
"summed_initial_probability_mass_outside_full_target": 17.380628072727674,
"summed_initial_full_target_nll": 42.72775317225145
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 140,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 46,
"actual_model_predictions": 127,
"actual_feedback_predictions": 63,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 50,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
9,
8,
7
],
"complete_function_curve_budgets_1_to_4": [
8,
17,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
140,
298,
410,
512
],
"summed_observed_prediction_ns": 1369659,
"summed_initial_probability_mass_outside_full_target": 23.975692659184034,
"summed_initial_full_target_nll": 44.29112537198986
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 246,
"initial_complete_functions": 14,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 39,
"actual_model_predictions": 116,
"actual_feedback_predictions": 52,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 57,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
14,
4,
7,
7
],
"complete_function_curve_budgets_1_to_4": [
14,
18,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
246,
303,
409,
512
],
"summed_observed_prediction_ns": 1326789,
"summed_initial_probability_mass_outside_full_target": 23.881407951430596,
"summed_initial_full_target_nll": 43.907869135967424
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 145,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 47,
"actual_model_predictions": 128,
"actual_feedback_predictions": 64,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 49,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
9,
7,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
17,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
145,
290,
400,
512
],
"summed_observed_prediction_ns": 1393536,
"summed_initial_probability_mass_outside_full_target": 24.000471580105337,
"summed_initial_full_target_nll": 44.40211283864889
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 129,
"initial_complete_functions": 6,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 50,
"actual_model_predictions": 137,
"actual_feedback_predictions": 73,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 46,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
6,
9,
10,
7
],
"complete_function_curve_budgets_1_to_4": [
6,
15,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
129,
268,
414,
512
],
"summed_observed_prediction_ns": 1496658,
"summed_initial_probability_mass_outside_full_target": 23.810924714855535,
"summed_initial_full_target_nll": 43.676219544727466
}
}
},
"joint-fp32": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2742,
"initial_complete_functions": 144,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 435,
"actual_model_predictions": 754,
"actual_feedback_predictions": 370,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 717,
"bilingual_initial_mask_disagreements": 174,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
144,
110,
65,
65
],
"complete_function_curve_budgets_1_to_4": [
144,
254,
319,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2742,
4250,
5266,
6144
],
"summed_observed_prediction_ns": 9034301,
"summed_initial_probability_mass_outside_full_target": 265.097424402833,
"summed_initial_full_target_nll": 525.706839299999
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 454,
"initial_complete_functions": 28,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 4,
"actual_model_predictions": 36,
"actual_feedback_predictions": 4,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 92,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
28,
4,
0,
0
],
"complete_function_curve_budgets_1_to_4": [
28,
32,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
454,
512,
512,
512
],
"summed_observed_prediction_ns": 397626,
"summed_initial_probability_mass_outside_full_target": 19.295918405056,
"summed_initial_full_target_nll": 42.05230999828956
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 283,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 24,
"actual_model_predictions": 56,
"actual_feedback_predictions": 24,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 72,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
8,
8,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
24,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
283,
399,
512,
512
],
"summed_observed_prediction_ns": 691996,
"summed_initial_probability_mass_outside_full_target": 19.75831100344658,
"summed_initial_full_target_nll": 44.354622241116715
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 197,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 61,
"actual_feedback_predictions": 29,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
11,
6,
3
],
"complete_function_curve_budgets_1_to_4": [
12,
23,
29,
32
],
"best_passing_case_curve_budgets_1_to_4": [
197,
372,
466,
512
],
"summed_observed_prediction_ns": 703751,
"summed_initial_probability_mass_outside_full_target": 22.962988913059235,
"summed_initial_full_target_nll": 40.60489804292393
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 135,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 34,
"actual_model_predictions": 64,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 62,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
16,
6,
2
],
"complete_function_curve_budgets_1_to_4": [
8,
24,
30,
32
],
"best_passing_case_curve_budgets_1_to_4": [
135,
387,
481,
512
],
"summed_observed_prediction_ns": 773042,
"summed_initial_probability_mass_outside_full_target": 23.989022821187973,
"summed_initial_full_target_nll": 45.035760070065955
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 148,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
148,
278,
396,
512
],
"summed_observed_prediction_ns": 860210,
"summed_initial_probability_mass_outside_full_target": 23.692762196063995,
"summed_initial_full_target_nll": 43.36107756314573
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 156,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 47,
"actual_model_predictions": 71,
"actual_feedback_predictions": 39,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 49,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
9,
7,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
17,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
156,
289,
396,
512
],
"summed_observed_prediction_ns": 981792,
"summed_initial_probability_mass_outside_full_target": 23.94161143898964,
"summed_initial_full_target_nll": 44.776263126191346
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 323,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 60,
"actual_feedback_predictions": 28,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
12,
4,
4
],
"complete_function_curve_budgets_1_to_4": [
12,
24,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
323,
401,
478,
512
],
"summed_observed_prediction_ns": 693500,
"summed_initial_probability_mass_outside_full_target": 17.63266098499298,
"summed_initial_full_target_nll": 43.67686643914
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 438,
"initial_complete_functions": 18,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 22,
"actual_model_predictions": 50,
"actual_feedback_predictions": 18,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 74,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
18,
10,
0,
4
],
"complete_function_curve_budgets_1_to_4": [
18,
28,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
438,
478,
478,
512
],
"summed_observed_prediction_ns": 592501,
"summed_initial_probability_mass_outside_full_target": 17.861520037055016,
"summed_initial_full_target_nll": 44.442449537424736
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 138,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 49,
"actual_model_predictions": 73,
"actual_feedback_predictions": 41,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 47,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
7,
9,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
15,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
138,
245,
401,
512
],
"summed_observed_prediction_ns": 863796,
"summed_initial_probability_mass_outside_full_target": 23.99233341217041,
"summed_initial_full_target_nll": 44.35710612136575
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 145,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 50,
"actual_model_predictions": 75,
"actual_feedback_predictions": 43,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 46,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
5,
12,
7
],
"complete_function_curve_budgets_1_to_4": [
8,
13,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
145,
238,
413,
512
],
"summed_observed_prediction_ns": 905173,
"summed_initial_probability_mass_outside_full_target": 23.991549998521805,
"summed_initial_full_target_nll": 44.39267490885772
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 178,
"initial_complete_functions": 10,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 47,
"actual_model_predictions": 67,
"actual_feedback_predictions": 35,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 49,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
10,
9,
1,
12
],
"complete_function_curve_budgets_1_to_4": [
10,
19,
20,
32
],
"best_passing_case_curve_budgets_1_to_4": [
178,
327,
345,
512
],
"summed_observed_prediction_ns": 770126,
"summed_initial_probability_mass_outside_full_target": 24.004550874233246,
"summed_initial_full_target_nll": 44.387888447202634
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 147,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 46,
"actual_model_predictions": 69,
"actual_feedback_predictions": 37,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 50,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
11,
4,
9
],
"complete_function_curve_budgets_1_to_4": [
8,
19,
23,
32
],
"best_passing_case_curve_budgets_1_to_4": [
147,
324,
388,
512
],
"summed_observed_prediction_ns": 800788,
"summed_initial_probability_mass_outside_full_target": 23.974194318056107,
"summed_initial_full_target_nll": 44.2649228042747
}
}
},
"joint-ptq_ternary": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2708,
"initial_complete_functions": 145,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 473,
"actual_model_predictions": 779,
"actual_feedback_predictions": 395,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 679,
"bilingual_initial_mask_disagreements": 44,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
145,
83,
78,
78
],
"complete_function_curve_budgets_1_to_4": [
145,
228,
306,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2708,
3960,
5066,
6144
],
"summed_observed_prediction_ns": 9318991,
"summed_initial_probability_mass_outside_full_target": 266.2105183303356,
"summed_initial_full_target_nll": 530.358279973884
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 311,
"initial_complete_functions": 18,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 34,
"actual_model_predictions": 58,
"actual_feedback_predictions": 26,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 62,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
18,
2,
4,
8
],
"complete_function_curve_budgets_1_to_4": [
18,
20,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
311,
339,
396,
512
],
"summed_observed_prediction_ns": 682167,
"summed_initial_probability_mass_outside_full_target": 19.47901126742363,
"summed_initial_full_target_nll": 43.485937632730774
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 283,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 33,
"actual_model_predictions": 62,
"actual_feedback_predictions": 30,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 63,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
2,
11,
3
],
"complete_function_curve_budgets_1_to_4": [
16,
18,
29,
32
],
"best_passing_case_curve_budgets_1_to_4": [
283,
311,
468,
512
],
"summed_observed_prediction_ns": 727585,
"summed_initial_probability_mass_outside_full_target": 19.622833400964737,
"summed_initial_full_target_nll": 44.30568677905417
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 261,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 39,
"actual_model_predictions": 64,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 57,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
0,
9,
7
],
"complete_function_curve_budgets_1_to_4": [
16,
16,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
261,
262,
406,
512
],
"summed_observed_prediction_ns": 830913,
"summed_initial_probability_mass_outside_full_target": 23.538529247045517,
"summed_initial_full_target_nll": 42.73027621232325
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 135,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 36,
"actual_model_predictions": 68,
"actual_feedback_predictions": 36,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 60,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
12,
12,
0
],
"complete_function_curve_budgets_1_to_4": [
8,
20,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
135,
325,
512,
512
],
"summed_observed_prediction_ns": 802709,
"summed_initial_probability_mass_outside_full_target": 23.978688687086105,
"summed_initial_full_target_nll": 44.63129659102326
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 186,
"initial_complete_functions": 10,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 50,
"actual_model_predictions": 70,
"actual_feedback_predictions": 38,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 46,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
10,
6,
4,
12
],
"complete_function_curve_budgets_1_to_4": [
10,
16,
20,
32
],
"best_passing_case_curve_budgets_1_to_4": [
186,
280,
341,
512
],
"summed_observed_prediction_ns": 844709,
"summed_initial_probability_mass_outside_full_target": 23.85689190030098,
"summed_initial_full_target_nll": 43.95067680696744
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 156,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 41,
"actual_model_predictions": 65,
"actual_feedback_predictions": 33,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 55,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
15,
1,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
23,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
156,
378,
396,
512
],
"summed_observed_prediction_ns": 763746,
"summed_initial_probability_mass_outside_full_target": 23.98713657259941,
"summed_initial_full_target_nll": 44.80060669726497
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 412,
"initial_complete_functions": 23,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 22,
"actual_model_predictions": 50,
"actual_feedback_predictions": 18,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 74,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
23,
0,
5,
4
],
"complete_function_curve_budgets_1_to_4": [
23,
23,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
412,
443,
481,
512
],
"summed_observed_prediction_ns": 579250,
"summed_initial_probability_mass_outside_full_target": 17.823746651411057,
"summed_initial_full_target_nll": 44.068231928590706
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 358,
"initial_complete_functions": 14,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 26,
"actual_model_predictions": 54,
"actual_feedback_predictions": 22,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 70,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
14,
14,
0,
4
],
"complete_function_curve_budgets_1_to_4": [
14,
28,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
358,
478,
478,
512
],
"summed_observed_prediction_ns": 634499,
"summed_initial_probability_mass_outside_full_target": 17.94053691625595,
"summed_initial_full_target_nll": 44.5215218634449
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 145,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
145,
283,
394,
512
],
"summed_observed_prediction_ns": 834547,
"summed_initial_probability_mass_outside_full_target": 23.999182522296906,
"summed_initial_full_target_nll": 44.49356924337915
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 145,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
145,
283,
394,
512
],
"summed_observed_prediction_ns": 903583,
"summed_initial_probability_mass_outside_full_target": 23.991287380456924,
"summed_initial_full_target_nll": 44.45797759965371
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 158,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
158,
289,
400,
512
],
"summed_observed_prediction_ns": 838456,
"summed_initial_probability_mass_outside_full_target": 23.999674290418625,
"summed_initial_full_target_nll": 44.46622133538837
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 158,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
158,
289,
400,
512
],
"summed_observed_prediction_ns": 876827,
"summed_initial_probability_mass_outside_full_target": 23.992999494075775,
"summed_initial_full_target_nll": 44.44627728406336
}
}
},
"joint-qat_ternary": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2423,
"initial_complete_functions": 124,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 479,
"actual_model_predictions": 792,
"actual_feedback_predictions": 408,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 673,
"bilingual_initial_mask_disagreements": 114,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
124,
112,
77,
71
],
"complete_function_curve_budgets_1_to_4": [
124,
236,
313,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2423,
3962,
5178,
6144
],
"summed_observed_prediction_ns": 9336370,
"summed_initial_probability_mass_outside_full_target": 264.97162987291813,
"summed_initial_full_target_nll": 525.6855548755702
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 283,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 56,
"actual_feedback_predictions": 24,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
8,
0,
8
],
"complete_function_curve_budgets_1_to_4": [
16,
24,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
283,
399,
399,
512
],
"summed_observed_prediction_ns": 660877,
"summed_initial_probability_mass_outside_full_target": 19.348411053419113,
"summed_initial_full_target_nll": 42.0239666770398
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 157,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 64,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
16,
8,
0
],
"complete_function_curve_budgets_1_to_4": [
8,
24,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
157,
399,
512,
512
],
"summed_observed_prediction_ns": 760127,
"summed_initial_probability_mass_outside_full_target": 19.84671440720558,
"summed_initial_full_target_nll": 43.96859888444768
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 197,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 38,
"actual_model_predictions": 64,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 58,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
8,
6,
6
],
"complete_function_curve_budgets_1_to_4": [
12,
20,
26,
32
],
"best_passing_case_curve_budgets_1_to_4": [
197,
328,
421,
512
],
"summed_observed_prediction_ns": 735461,
"summed_initial_probability_mass_outside_full_target": 22.927854120731354,
"summed_initial_full_target_nll": 40.61118979666142
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 136,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 33,
"actual_model_predictions": 64,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 63,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
16,
7,
1
],
"complete_function_curve_budgets_1_to_4": [
8,
24,
31,
32
],
"best_passing_case_curve_budgets_1_to_4": [
136,
391,
496,
512
],
"summed_observed_prediction_ns": 736664,
"summed_initial_probability_mass_outside_full_target": 23.95988169312477,
"summed_initial_full_target_nll": 45.189708459635185
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 148,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 49,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 47,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
7,
9
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
23,
32
],
"best_passing_case_curve_budgets_1_to_4": [
148,
278,
381,
512
],
"summed_observed_prediction_ns": 828712,
"summed_initial_probability_mass_outside_full_target": 23.709769248962402,
"summed_initial_full_target_nll": 43.43819566358455
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 154,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 47,
"actual_model_predictions": 71,
"actual_feedback_predictions": 39,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 49,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
9,
7,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
17,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
154,
288,
396,
512
],
"summed_observed_prediction_ns": 827622,
"summed_initial_probability_mass_outside_full_target": 23.902148693799973,
"summed_initial_full_target_nll": 45.18439960906941
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 353,
"initial_complete_functions": 14,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 30,
"actual_model_predictions": 58,
"actual_feedback_predictions": 26,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 66,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
14,
10,
4,
4
],
"complete_function_curve_budgets_1_to_4": [
14,
24,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
353,
401,
478,
512
],
"summed_observed_prediction_ns": 712794,
"summed_initial_probability_mass_outside_full_target": 17.664475470781326,
"summed_initial_full_target_nll": 43.719917149763944
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 357,
"initial_complete_functions": 14,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 26,
"actual_model_predictions": 54,
"actual_feedback_predictions": 22,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 70,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
14,
14,
0,
4
],
"complete_function_curve_budgets_1_to_4": [
14,
28,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
357,
468,
478,
512
],
"summed_observed_prediction_ns": 653075,
"summed_initial_probability_mass_outside_full_target": 17.800772562623024,
"summed_initial_full_target_nll": 44.59855760982592
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 138,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
138,
258,
401,
512
],
"summed_observed_prediction_ns": 851080,
"summed_initial_probability_mass_outside_full_target": 23.97306102514267,
"summed_initial_full_target_nll": 44.349513222215805
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 138,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 70,
"actual_feedback_predictions": 38,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
10,
4,
10
],
"complete_function_curve_budgets_1_to_4": [
8,
18,
22,
32
],
"best_passing_case_curve_budgets_1_to_4": [
138,
290,
373,
512
],
"summed_observed_prediction_ns": 831960,
"summed_initial_probability_mass_outside_full_target": 23.9512600004673,
"summed_initial_full_target_nll": 44.19247254222447
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 145,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 50,
"actual_model_predictions": 75,
"actual_feedback_predictions": 43,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 46,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
5,
12,
7
],
"complete_function_curve_budgets_1_to_4": [
8,
13,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
145,
239,
415,
512
],
"summed_observed_prediction_ns": 875410,
"summed_initial_probability_mass_outside_full_target": 23.991551250219345,
"summed_initial_full_target_nll": 44.43531872986863
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 217,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 46,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 50,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
0,
14,
6
],
"complete_function_curve_budgets_1_to_4": [
12,
12,
26,
32
],
"best_passing_case_curve_budgets_1_to_4": [
217,
223,
428,
512
],
"summed_observed_prediction_ns": 862588,
"summed_initial_probability_mass_outside_full_target": 23.89573034644127,
"summed_initial_full_target_nll": 43.97371653123351
}
}
},
"offline": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2318,
"initial_complete_functions": 116,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 516,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 636,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
116,
100,
88,
80
],
"complete_function_curve_budgets_1_to_4": [
116,
216,
304,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2318,
3808,
5038,
6144
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 268,
"summed_initial_full_target_nll": 7957.734081387455
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 222,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
12,
4,
4
],
"complete_function_curve_budgets_1_to_4": [
12,
24,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
222,
396,
455,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 20,
"summed_initial_full_target_nll": 663.1445067822851
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 222,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
12,
4,
4
],
"complete_function_curve_budgets_1_to_4": [
12,
24,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
222,
396,
455,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 20,
"summed_initial_full_target_nll": 663.1445067822851
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 134,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
134,
265,
390,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 134,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
134,
265,
390,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 155,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
155,
276,
397,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 155,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
155,
276,
397,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 354,
"initial_complete_functions": 14,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 34,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 62,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
14,
6,
8,
4
],
"complete_function_curve_budgets_1_to_4": [
14,
20,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
354,
416,
478,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 18,
"summed_initial_full_target_nll": 663.1445067822847
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 354,
"initial_complete_functions": 14,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 34,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 62,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
14,
6,
8,
4
],
"complete_function_curve_budgets_1_to_4": [
14,
20,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
354,
416,
478,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 18,
"summed_initial_full_target_nll": 663.1445067822847
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 141,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
141,
273,
398,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 141,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
141,
273,
398,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 153,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
153,
278,
401,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 153,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
153,
278,
401,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
}
}
}
},
"dataset_sha256": "2d1c9888a648d590e77253666b6e5ec848d375e47a3c99daf29dbde8bcbc7383",
"default_model_promoted": false,
"development": {
"independent-fp32": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2596,
"initial_complete_functions": 124,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 448,
"actual_model_predictions": 1447,
"actual_feedback_predictions": 679,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 704,
"bilingual_initial_mask_disagreements": 182,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
124,
130,
72,
58
],
"complete_function_curve_budgets_1_to_4": [
124,
254,
326,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2596,
4288,
5356,
6144
],
"summed_observed_prediction_ns": 13756082,
"summed_initial_probability_mass_outside_full_target": 270.2369040239579,
"summed_initial_full_target_nll": 529.3969786973563
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 282,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 36,
"actual_model_predictions": 120,
"actual_feedback_predictions": 56,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 60,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
4,
4,
8
],
"complete_function_curve_budgets_1_to_4": [
16,
20,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
282,
335,
396,
512
],
"summed_observed_prediction_ns": 1089122,
"summed_initial_probability_mass_outside_full_target": 19.454412209402626,
"summed_initial_full_target_nll": 43.344522632874806
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 282,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 16,
"actual_model_predictions": 96,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 80,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
16,
0,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
32,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
282,
512,
512,
512
],
"summed_observed_prediction_ns": 864500,
"summed_initial_probability_mass_outside_full_target": 20.139105324603825,
"summed_initial_full_target_nll": 45.59576241078689
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 150,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 47,
"actual_model_predictions": 127,
"actual_feedback_predictions": 63,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 49,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
9,
7,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
17,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
150,
284,
395,
512
],
"summed_observed_prediction_ns": 1173502,
"summed_initial_probability_mass_outside_full_target": 23.584299142719313,
"summed_initial_full_target_nll": 43.05492018665735
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 216,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 27,
"actual_model_predictions": 114,
"actual_feedback_predictions": 50,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 69,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
13,
7,
0
],
"complete_function_curve_budgets_1_to_4": [
12,
25,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
216,
406,
512,
512
],
"summed_observed_prediction_ns": 1082830,
"summed_initial_probability_mass_outside_full_target": 24.03600345658598,
"summed_initial_full_target_nll": 44.86495176361099
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 152,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 46,
"actual_model_predictions": 131,
"actual_feedback_predictions": 67,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 50,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
9,
8,
7
],
"complete_function_curve_budgets_1_to_4": [
8,
17,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
152,
295,
413,
512
],
"summed_observed_prediction_ns": 1207625,
"summed_initial_probability_mass_outside_full_target": 23.52249151649616,
"summed_initial_full_target_nll": 42.79543009501006
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 166,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 33,
"actual_model_predictions": 125,
"actual_feedback_predictions": 61,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 63,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
15,
9,
0
],
"complete_function_curve_budgets_1_to_4": [
8,
23,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
166,
378,
512,
512
],
"summed_observed_prediction_ns": 1215378,
"summed_initial_probability_mass_outside_full_target": 24.195120364891046,
"summed_initial_full_target_nll": 45.297945959234454
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 349,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 112,
"actual_feedback_predictions": 48,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
12,
4,
4
],
"complete_function_curve_budgets_1_to_4": [
12,
24,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
349,
398,
478,
512
],
"summed_observed_prediction_ns": 1119580,
"summed_initial_probability_mass_outside_full_target": 19.303478971726875,
"summed_initial_full_target_nll": 42.89452533331658
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 436,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 22,
"actual_model_predictions": 108,
"actual_feedback_predictions": 44,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 74,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
10,
6,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
26,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
436,
462,
512,
512
],
"summed_observed_prediction_ns": 974000,
"summed_initial_probability_mass_outside_full_target": 20.172023861815838,
"summed_initial_full_target_nll": 44.67739308860217
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 84,
"initial_complete_functions": 4,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 51,
"actual_model_predictions": 135,
"actual_feedback_predictions": 71,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 45,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
4,
13,
7,
8
],
"complete_function_curve_budgets_1_to_4": [
4,
17,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
84,
305,
404,
512
],
"summed_observed_prediction_ns": 1296039,
"summed_initial_probability_mass_outside_full_target": 23.991596790878955,
"summed_initial_full_target_nll": 44.35048585661826
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 211,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 43,
"actual_model_predictions": 119,
"actual_feedback_predictions": 55,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 53,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
5,
7,
8
],
"complete_function_curve_budgets_1_to_4": [
12,
17,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
211,
288,
401,
512
],
"summed_observed_prediction_ns": 1131835,
"summed_initial_probability_mass_outside_full_target": 23.921435067123383,
"summed_initial_full_target_nll": 44.06931229154658
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 108,
"initial_complete_functions": 4,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 50,
"actual_model_predictions": 135,
"actual_feedback_predictions": 71,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 46,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
4,
13,
8,
7
],
"complete_function_curve_budgets_1_to_4": [
4,
17,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
108,
291,
417,
512
],
"summed_observed_prediction_ns": 1311341,
"summed_initial_probability_mass_outside_full_target": 24.00850048699855,
"summed_initial_full_target_nll": 44.404286398846736
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 160,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 45,
"actual_model_predictions": 125,
"actual_feedback_predictions": 61,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 51,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
11,
5,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
19,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
160,
334,
404,
512
],
"summed_observed_prediction_ns": 1290330,
"summed_initial_probability_mass_outside_full_target": 23.908436830714706,
"summed_initial_full_target_nll": 44.04744268025131
}
}
},
"independent-ptq_ternary": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2307,
"initial_complete_functions": 108,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 479,
"actual_model_predictions": 1470,
"actual_feedback_predictions": 702,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 673,
"bilingual_initial_mask_disagreements": 136,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
108,
131,
87,
58
],
"complete_function_curve_budgets_1_to_4": [
108,
239,
326,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2307,
4105,
5380,
6144
],
"summed_observed_prediction_ns": 15919801,
"summed_initial_probability_mass_outside_full_target": 270.8211417307867,
"summed_initial_full_target_nll": 531.008970892193
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 282,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 21,
"actual_model_predictions": 101,
"actual_feedback_predictions": 37,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 75,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
11,
5,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
27,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
282,
438,
512,
512
],
"summed_observed_prediction_ns": 1020247,
"summed_initial_probability_mass_outside_full_target": 19.58995282464997,
"summed_initial_full_target_nll": 43.94805563644665
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 156,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 31,
"actual_model_predictions": 119,
"actual_feedback_predictions": 55,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 65,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
17,
7,
0
],
"complete_function_curve_budgets_1_to_4": [
8,
25,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
156,
415,
512,
512
],
"summed_observed_prediction_ns": 1281747,
"summed_initial_probability_mass_outside_full_target": 20.12548944983616,
"summed_initial_full_target_nll": 44.91999736054559
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 150,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 47,
"actual_model_predictions": 127,
"actual_feedback_predictions": 63,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 49,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
9,
7,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
17,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
150,
281,
390,
512
],
"summed_observed_prediction_ns": 1391378,
"summed_initial_probability_mass_outside_full_target": 23.713772109858684,
"summed_initial_full_target_nll": 43.64070089090548
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 155,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 37,
"actual_model_predictions": 125,
"actual_feedback_predictions": 61,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 59,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
11,
13,
0
],
"complete_function_curve_budgets_1_to_4": [
8,
19,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
155,
314,
512,
512
],
"summed_observed_prediction_ns": 1313379,
"summed_initial_probability_mass_outside_full_target": 24.03457078098702,
"summed_initial_full_target_nll": 44.59113540272631
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 89,
"initial_complete_functions": 4,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 47,
"actual_model_predictions": 135,
"actual_feedback_predictions": 71,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 49,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
4,
13,
11,
4
],
"complete_function_curve_budgets_1_to_4": [
4,
17,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
89,
292,
458,
512
],
"summed_observed_prediction_ns": 1433665,
"summed_initial_probability_mass_outside_full_target": 23.678709436267663,
"summed_initial_full_target_nll": 43.464071335228425
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 215,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 40,
"actual_model_predictions": 118,
"actual_feedback_predictions": 54,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 56,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
6,
8,
6
],
"complete_function_curve_budgets_1_to_4": [
12,
18,
26,
32
],
"best_passing_case_curve_budgets_1_to_4": [
215,
308,
427,
512
],
"summed_observed_prediction_ns": 1248920,
"summed_initial_probability_mass_outside_full_target": 24.140005747662965,
"summed_initial_full_target_nll": 45.02279083272974
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 349,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 31,
"actual_model_predictions": 111,
"actual_feedback_predictions": 47,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 65,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
13,
3,
4
],
"complete_function_curve_budgets_1_to_4": [
12,
25,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
349,
448,
478,
512
],
"summed_observed_prediction_ns": 1179077,
"summed_initial_probability_mass_outside_full_target": 19.5558538160392,
"summed_initial_full_target_nll": 43.60809592312091
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 349,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 29,
"actual_model_predictions": 109,
"actual_feedback_predictions": 45,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 67,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
15,
1,
4
],
"complete_function_curve_budgets_1_to_4": [
12,
27,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
349,
460,
478,
512
],
"summed_observed_prediction_ns": 1185290,
"summed_initial_probability_mass_outside_full_target": 20.059972514073458,
"summed_initial_full_target_nll": 44.40973699463653
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 84,
"initial_complete_functions": 4,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 52,
"actual_model_predictions": 136,
"actual_feedback_predictions": 72,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 44,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
4,
12,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
4,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
84,
289,
403,
512
],
"summed_observed_prediction_ns": 1543422,
"summed_initial_probability_mass_outside_full_target": 24.000360483152786,
"summed_initial_full_target_nll": 44.45233420054842
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 155,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 128,
"actual_feedback_predictions": 64,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
155,
288,
401,
512
],
"summed_observed_prediction_ns": 1420295,
"summed_initial_probability_mass_outside_full_target": 23.967537565014595,
"summed_initial_full_target_nll": 44.29192324151203
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 95,
"initial_complete_functions": 4,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 52,
"actual_model_predictions": 140,
"actual_feedback_predictions": 76,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 44,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
4,
13,
6,
9
],
"complete_function_curve_budgets_1_to_4": [
4,
17,
23,
32
],
"best_passing_case_curve_budgets_1_to_4": [
95,
294,
391,
512
],
"summed_observed_prediction_ns": 1565501,
"summed_initial_probability_mass_outside_full_target": 23.996031074746778,
"summed_initial_full_target_nll": 44.373301873671785
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 228,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 44,
"actual_model_predictions": 121,
"actual_feedback_predictions": 57,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 52,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
3,
10,
7
],
"complete_function_curve_budgets_1_to_4": [
12,
15,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
228,
278,
418,
512
],
"summed_observed_prediction_ns": 1336880,
"summed_initial_probability_mass_outside_full_target": 23.958885928497683,
"summed_initial_full_target_nll": 44.286827200121316
}
}
},
"independent-qat_ternary": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2509,
"initial_complete_functions": 120,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 455,
"actual_model_predictions": 1453,
"actual_feedback_predictions": 685,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 697,
"bilingual_initial_mask_disagreements": 182,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
120,
131,
75,
58
],
"complete_function_curve_budgets_1_to_4": [
120,
251,
326,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2509,
4253,
5367,
6144
],
"summed_observed_prediction_ns": 15927855,
"summed_initial_probability_mass_outside_full_target": 270.4540155270739,
"summed_initial_full_target_nll": 529.3311030850092
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 282,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 33,
"actual_model_predictions": 110,
"actual_feedback_predictions": 46,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 63,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
7,
1,
8
],
"complete_function_curve_budgets_1_to_4": [
16,
23,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
282,
381,
396,
512
],
"summed_observed_prediction_ns": 1141291,
"summed_initial_probability_mass_outside_full_target": 19.5702818550002,
"summed_initial_full_target_nll": 43.61249070486638
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 282,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 16,
"actual_model_predictions": 96,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 80,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
16,
0,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
32,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
282,
512,
512,
512
],
"summed_observed_prediction_ns": 978378,
"summed_initial_probability_mass_outside_full_target": 20.098246110239913,
"summed_initial_full_target_nll": 45.20473649221639
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 150,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 46,
"actual_model_predictions": 127,
"actual_feedback_predictions": 63,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 50,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
10,
6,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
18,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
150,
299,
392,
512
],
"summed_observed_prediction_ns": 1393834,
"summed_initial_probability_mass_outside_full_target": 23.64181559662097,
"summed_initial_full_target_nll": 43.37011118107379
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 216,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 27,
"actual_model_predictions": 114,
"actual_feedback_predictions": 50,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 69,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
13,
7,
0
],
"complete_function_curve_budgets_1_to_4": [
12,
25,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
216,
406,
512,
512
],
"summed_observed_prediction_ns": 1331119,
"summed_initial_probability_mass_outside_full_target": 23.973971136092583,
"summed_initial_full_target_nll": 44.46921377369318
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 152,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 47,
"actual_model_predictions": 130,
"actual_feedback_predictions": 66,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 49,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
9,
7,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
17,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
152,
295,
398,
512
],
"summed_observed_prediction_ns": 1407759,
"summed_initial_probability_mass_outside_full_target": 23.646611086088388,
"summed_initial_full_target_nll": 43.11466373324452
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 166,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 124,
"actual_feedback_predictions": 60,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
16,
8,
0
],
"complete_function_curve_budgets_1_to_4": [
8,
24,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
166,
394,
512,
512
],
"summed_observed_prediction_ns": 1330336,
"summed_initial_probability_mass_outside_full_target": 24.14050346223522,
"summed_initial_full_target_nll": 45.10495157595851
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 262,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 43,
"actual_model_predictions": 128,
"actual_feedback_predictions": 64,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 53,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
12,
5,
7
],
"complete_function_curve_budgets_1_to_4": [
8,
20,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
262,
364,
454,
512
],
"summed_observed_prediction_ns": 1441209,
"summed_initial_probability_mass_outside_full_target": 19.517105786555494,
"summed_initial_full_target_nll": 43.2892556697893
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 436,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 20,
"actual_model_predictions": 104,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 76,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
12,
4,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
28,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
436,
478,
512,
512
],
"summed_observed_prediction_ns": 1095418,
"summed_initial_probability_mass_outside_full_target": 20.015601767231146,
"summed_initial_full_target_nll": 44.16782908120292
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 84,
"initial_complete_functions": 4,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 52,
"actual_model_predictions": 137,
"actual_feedback_predictions": 73,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 44,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
4,
11,
10,
7
],
"complete_function_curve_budgets_1_to_4": [
4,
15,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
84,
269,
417,
512
],
"summed_observed_prediction_ns": 1527051,
"summed_initial_probability_mass_outside_full_target": 23.994678123135266,
"summed_initial_full_target_nll": 44.37544120572716
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 211,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 42,
"actual_model_predictions": 120,
"actual_feedback_predictions": 56,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 54,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
4,
10,
6
],
"complete_function_curve_budgets_1_to_4": [
12,
16,
26,
32
],
"best_passing_case_curve_budgets_1_to_4": [
211,
272,
425,
512
],
"summed_observed_prediction_ns": 1351379,
"summed_initial_probability_mass_outside_full_target": 23.941040831911284,
"summed_initial_full_target_nll": 44.154172901818406
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 108,
"initial_complete_functions": 4,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 50,
"actual_model_predictions": 136,
"actual_feedback_predictions": 72,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 46,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
4,
12,
10,
6
],
"complete_function_curve_budgets_1_to_4": [
4,
16,
26,
32
],
"best_passing_case_curve_budgets_1_to_4": [
108,
277,
432,
512
],
"summed_observed_prediction_ns": 1529791,
"summed_initial_probability_mass_outside_full_target": 24.004105441701263,
"summed_initial_full_target_nll": 44.39690641397221
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 160,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 47,
"actual_model_predictions": 127,
"actual_feedback_predictions": 63,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 49,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
9,
7,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
17,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
160,
306,
405,
512
],
"summed_observed_prediction_ns": 1400290,
"summed_initial_probability_mass_outside_full_target": 23.910054330261914,
"summed_initial_full_target_nll": 44.07133035144721
}
}
},
"joint-fp32": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2639,
"initial_complete_functions": 128,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 455,
"actual_model_predictions": 777,
"actual_feedback_predictions": 393,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 697,
"bilingual_initial_mask_disagreements": 180,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
128,
119,
75,
62
],
"complete_function_curve_budgets_1_to_4": [
128,
247,
322,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2639,
4171,
5291,
6144
],
"summed_observed_prediction_ns": 9500791,
"summed_initial_probability_mass_outside_full_target": 271.0070815384388,
"summed_initial_full_target_nll": 530.778929149537
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 395,
"initial_complete_functions": 24,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 14,
"actual_model_predictions": 46,
"actual_feedback_predictions": 14,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 82,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
24,
2,
6,
0
],
"complete_function_curve_budgets_1_to_4": [
24,
26,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
395,
425,
512,
512
],
"summed_observed_prediction_ns": 529042,
"summed_initial_probability_mass_outside_full_target": 19.70918470621109,
"summed_initial_full_target_nll": 43.584187896788585
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 282,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 24,
"actual_model_predictions": 56,
"actual_feedback_predictions": 24,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 72,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
8,
8,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
24,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
282,
399,
512,
512
],
"summed_observed_prediction_ns": 687464,
"summed_initial_probability_mass_outside_full_target": 19.967933505773544,
"summed_initial_full_target_nll": 44.87374996803259
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 146,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
146,
262,
385,
512
],
"summed_observed_prediction_ns": 851002,
"summed_initial_probability_mass_outside_full_target": 23.63910436630249,
"summed_initial_full_target_nll": 43.25037554821077
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 140,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 37,
"actual_model_predictions": 64,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 59,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
16,
3,
5
],
"complete_function_curve_budgets_1_to_4": [
8,
24,
27,
32
],
"best_passing_case_curve_budgets_1_to_4": [
140,
388,
438,
512
],
"summed_observed_prediction_ns": 768960,
"summed_initial_probability_mass_outside_full_target": 24.081539511680603,
"summed_initial_full_target_nll": 44.843781387203016
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 121,
"initial_complete_functions": 6,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 54,
"actual_model_predictions": 76,
"actual_feedback_predictions": 44,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 42,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
6,
8,
8,
10
],
"complete_function_curve_budgets_1_to_4": [
6,
14,
22,
32
],
"best_passing_case_curve_budgets_1_to_4": [
121,
248,
363,
512
],
"summed_observed_prediction_ns": 903829,
"summed_initial_probability_mass_outside_full_target": 23.72674262523651,
"summed_initial_full_target_nll": 43.436827578169385
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 166,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 64,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
16,
8,
0
],
"complete_function_curve_budgets_1_to_4": [
8,
24,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
166,
394,
512,
512
],
"summed_observed_prediction_ns": 748539,
"summed_initial_probability_mass_outside_full_target": 24.1494023501873,
"summed_initial_full_target_nll": 45.07644758544018
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 349,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 60,
"actual_feedback_predictions": 28,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
12,
4,
4
],
"complete_function_curve_budgets_1_to_4": [
12,
24,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
349,
398,
478,
512
],
"summed_observed_prediction_ns": 745418,
"summed_initial_probability_mass_outside_full_target": 19.61037689447403,
"summed_initial_full_target_nll": 43.479983873128
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 436,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 20,
"actual_model_predictions": 52,
"actual_feedback_predictions": 20,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 76,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
12,
4,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
28,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
436,
478,
512,
512
],
"summed_observed_prediction_ns": 629457,
"summed_initial_probability_mass_outside_full_target": 20.16948327422142,
"summed_initial_full_target_nll": 44.838219134947124
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 126,
"initial_complete_functions": 6,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 50,
"actual_model_predictions": 74,
"actual_feedback_predictions": 42,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 46,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
6,
10,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
6,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
126,
281,
406,
512
],
"summed_observed_prediction_ns": 901786,
"summed_initial_probability_mass_outside_full_target": 23.999244272708893,
"summed_initial_full_target_nll": 44.370327581345634
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 189,
"initial_complete_functions": 10,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 42,
"actual_model_predictions": 67,
"actual_feedback_predictions": 35,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 54,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
10,
9,
6,
7
],
"complete_function_curve_budgets_1_to_4": [
10,
19,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
189,
331,
417,
512
],
"summed_observed_prediction_ns": 958752,
"summed_initial_probability_mass_outside_full_target": 23.98909765481949,
"summed_initial_full_target_nll": 44.366243987301765
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 129,
"initial_complete_functions": 6,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 54,
"actual_model_predictions": 74,
"actual_feedback_predictions": 42,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 42,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
6,
10,
4,
12
],
"complete_function_curve_budgets_1_to_4": [
6,
16,
20,
32
],
"best_passing_case_curve_budgets_1_to_4": [
129,
291,
352,
512
],
"summed_observed_prediction_ns": 921792,
"summed_initial_probability_mass_outside_full_target": 23.994300425052643,
"summed_initial_full_target_nll": 44.351964492171916
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 160,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
160,
276,
404,
512
],
"summed_observed_prediction_ns": 854750,
"summed_initial_probability_mass_outside_full_target": 23.970671951770782,
"summed_initial_full_target_nll": 44.3068201167975
}
}
},
"joint-ptq_ternary": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2635,
"initial_complete_functions": 128,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 485,
"actual_model_predictions": 808,
"actual_feedback_predictions": 424,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 667,
"bilingual_initial_mask_disagreements": 24,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
128,
88,
107,
61
],
"complete_function_curve_budgets_1_to_4": [
128,
216,
323,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2635,
3927,
5306,
6144
],
"summed_observed_prediction_ns": 9913474,
"summed_initial_probability_mass_outside_full_target": 271.1784009039402,
"summed_initial_full_target_nll": 533.2008575664042
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 282,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 40,
"actual_model_predictions": 64,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 56,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
0,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
16,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
282,
282,
396,
512
],
"summed_observed_prediction_ns": 740500,
"summed_initial_probability_mass_outside_full_target": 19.624286621809006,
"summed_initial_full_target_nll": 44.11965010645702
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 282,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 28,
"actual_model_predictions": 60,
"actual_feedback_predictions": 28,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 68,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
4,
12,
0
],
"complete_function_curve_budgets_1_to_4": [
16,
20,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
282,
342,
512,
512
],
"summed_observed_prediction_ns": 719296,
"summed_initial_probability_mass_outside_full_target": 19.772160291671753,
"summed_initial_full_target_nll": 44.774355854440785
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 265,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 39,
"actual_model_predictions": 64,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 57,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
0,
9,
7
],
"complete_function_curve_budgets_1_to_4": [
16,
16,
25,
32
],
"best_passing_case_curve_budgets_1_to_4": [
265,
277,
403,
512
],
"summed_observed_prediction_ns": 774707,
"summed_initial_probability_mass_outside_full_target": 23.848025798797607,
"summed_initial_full_target_nll": 43.93552281976686
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 140,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 36,
"actual_model_predictions": 68,
"actual_feedback_predictions": 36,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 60,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
12,
12,
0
],
"complete_function_curve_budgets_1_to_4": [
8,
20,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
140,
339,
512,
512
],
"summed_observed_prediction_ns": 828751,
"summed_initial_probability_mass_outside_full_target": 24.031789749860764,
"summed_initial_full_target_nll": 44.65230342003479
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 158,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
158,
281,
398,
512
],
"summed_observed_prediction_ns": 843124,
"summed_initial_probability_mass_outside_full_target": 23.897545754909515,
"summed_initial_full_target_nll": 44.09979402983785
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 158,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 35,
"actual_model_predictions": 67,
"actual_feedback_predictions": 35,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 61,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
13,
11,
0
],
"complete_function_curve_budgets_1_to_4": [
8,
21,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
158,
349,
512,
512
],
"summed_observed_prediction_ns": 835251,
"summed_initial_probability_mass_outside_full_target": 24.07825180888176,
"summed_initial_full_target_nll": 44.875306904286916
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 353,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 37,
"actual_model_predictions": 65,
"actual_feedback_predictions": 33,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 59,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
7,
9,
4
],
"complete_function_curve_budgets_1_to_4": [
12,
19,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
353,
439,
478,
512
],
"summed_observed_prediction_ns": 777416,
"summed_initial_probability_mass_outside_full_target": 19.853339344263077,
"summed_initial_full_target_nll": 44.14603806517079
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 353,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 28,
"actual_model_predictions": 60,
"actual_feedback_predictions": 28,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 68,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
12,
8,
0
],
"complete_function_curve_budgets_1_to_4": [
12,
24,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
353,
444,
512,
512
],
"summed_observed_prediction_ns": 757420,
"summed_initial_probability_mass_outside_full_target": 20.084148108959198,
"summed_initial_full_target_nll": 44.71109173647279
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 158,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
158,
292,
401,
512
],
"summed_observed_prediction_ns": 870631,
"summed_initial_probability_mass_outside_full_target": 24.00152823328972,
"summed_initial_full_target_nll": 44.495768449591935
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 158,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
158,
292,
401,
512
],
"summed_observed_prediction_ns": 907624,
"summed_initial_probability_mass_outside_full_target": 23.994916886091232,
"summed_initial_full_target_nll": 44.47724179397162
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 164,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
164,
296,
404,
512
],
"summed_observed_prediction_ns": 873294,
"summed_initial_probability_mass_outside_full_target": 24.00063782930374,
"summed_initial_full_target_nll": 44.46258006811767
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 164,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 50,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 46,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
6,
10
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
22,
32
],
"best_passing_case_curve_budgets_1_to_4": [
164,
294,
377,
512
],
"summed_observed_prediction_ns": 985460,
"summed_initial_probability_mass_outside_full_target": 23.99177047610283,
"summed_initial_full_target_nll": 44.45120431825544
}
}
},
"joint-qat_ternary": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2394,
"initial_complete_functions": 114,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 469,
"actual_model_predictions": 792,
"actual_feedback_predictions": 408,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 683,
"bilingual_initial_mask_disagreements": 132,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
114,
132,
77,
61
],
"complete_function_curve_budgets_1_to_4": [
114,
246,
323,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2394,
4132,
5313,
6144
],
"summed_observed_prediction_ns": 9659504,
"summed_initial_probability_mass_outside_full_target": 271.5664424300194,
"summed_initial_full_target_nll": 531.9454298586631
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 282,
"initial_complete_functions": 16,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 30,
"actual_model_predictions": 56,
"actual_feedback_predictions": 24,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 66,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
16,
8,
2,
6
],
"complete_function_curve_budgets_1_to_4": [
16,
24,
26,
32
],
"best_passing_case_curve_budgets_1_to_4": [
282,
400,
425,
512
],
"summed_observed_prediction_ns": 661082,
"summed_initial_probability_mass_outside_full_target": 19.756559431552887,
"summed_initial_full_target_nll": 43.62048001899949
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 156,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 64,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
16,
8,
0
],
"complete_function_curve_budgets_1_to_4": [
8,
24,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
156,
400,
512,
512
],
"summed_observed_prediction_ns": 811495,
"summed_initial_probability_mass_outside_full_target": 20.210161596536636,
"summed_initial_full_target_nll": 45.039872947373944
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 150,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
150,
268,
391,
512
],
"summed_observed_prediction_ns": 860958,
"summed_initial_probability_mass_outside_full_target": 23.672243654727936,
"summed_initial_full_target_nll": 43.603786318080616
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 155,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 35,
"actual_model_predictions": 64,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 61,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
16,
5,
3
],
"complete_function_curve_budgets_1_to_4": [
8,
24,
29,
32
],
"best_passing_case_curve_budgets_1_to_4": [
155,
391,
465,
512
],
"summed_observed_prediction_ns": 781752,
"summed_initial_probability_mass_outside_full_target": 24.098874390125275,
"summed_initial_full_target_nll": 44.84797966183368
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 152,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
152,
282,
398,
512
],
"summed_observed_prediction_ns": 854839,
"summed_initial_probability_mass_outside_full_target": 23.76179540157318,
"summed_initial_full_target_nll": 43.50203847362654
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 162,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 64,
"actual_feedback_predictions": 32,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
16,
8,
0
],
"complete_function_curve_budgets_1_to_4": [
8,
24,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
162,
396,
512,
512
],
"summed_observed_prediction_ns": 782831,
"summed_initial_probability_mass_outside_full_target": 24.226962327957153,
"summed_initial_full_target_nll": 45.36419512755738
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 358,
"initial_complete_functions": 14,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 26,
"actual_model_predictions": 54,
"actual_feedback_predictions": 22,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 70,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
14,
14,
0,
4
],
"complete_function_curve_budgets_1_to_4": [
14,
28,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
358,
453,
478,
512
],
"summed_observed_prediction_ns": 629581,
"summed_initial_probability_mass_outside_full_target": 19.677556931972504,
"summed_initial_full_target_nll": 43.65170114181307
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 353,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 24,
"actual_model_predictions": 56,
"actual_feedback_predictions": 24,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 72,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
16,
4,
0
],
"complete_function_curve_budgets_1_to_4": [
12,
28,
32,
32
],
"best_passing_case_curve_budgets_1_to_4": [
353,
453,
512,
512
],
"summed_observed_prediction_ns": 703040,
"summed_initial_probability_mass_outside_full_target": 20.269226163625717,
"summed_initial_full_target_nll": 45.06549546193283
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 153,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
153,
269,
406,
512
],
"summed_observed_prediction_ns": 890833,
"summed_initial_probability_mass_outside_full_target": 23.9886714220047,
"summed_initial_full_target_nll": 44.38655418957031
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 153,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 46,
"actual_model_predictions": 70,
"actual_feedback_predictions": 38,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 50,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
10,
6,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
18,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
153,
302,
406,
512
],
"summed_observed_prediction_ns": 864915,
"summed_initial_probability_mass_outside_full_target": 23.983598321676254,
"summed_initial_full_target_nll": 44.31255996050225
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 160,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 52,
"actual_model_predictions": 76,
"actual_feedback_predictions": 44,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 44,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
4,
12,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
12,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
160,
235,
404,
512
],
"summed_observed_prediction_ns": 930254,
"summed_initial_probability_mass_outside_full_target": 23.9864062666893,
"summed_initial_full_target_nll": 44.38171781258787
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 160,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 72,
"actual_feedback_predictions": 40,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
160,
283,
404,
512
],
"summed_observed_prediction_ns": 887924,
"summed_initial_probability_mass_outside_full_target": 23.934386521577835,
"summed_initial_full_target_nll": 44.16904874478554
}
}
},
"offline": {
"total": {
"function_views": 384,
"ordered_finite_cases": 6144,
"initial_cases_passed": 2394,
"initial_complete_functions": 112,
"complete_functions_by_budget_four": 384,
"additional_candidate_attempts": 520,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 632,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 192,
"minimum_complete_budget_histogram_1_to_4": [
112,
104,
88,
80
],
"complete_function_curve_budgets_1_to_4": [
112,
216,
304,
384
],
"best_passing_case_curve_budgets_1_to_4": [
2394,
3854,
5052,
6144
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 272,
"summed_initial_full_target_nll": 7957.734081387455
},
"family_language": {
"assignment_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 221,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
12,
4,
4
],
"complete_function_curve_budgets_1_to_4": [
12,
24,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
221,
397,
456,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 20,
"summed_initial_full_target_nll": 663.1445067822851
},
"assignment_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 221,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 32,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 64,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
12,
4,
4
],
"complete_function_curve_budgets_1_to_4": [
12,
24,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
221,
397,
456,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 20,
"summed_initial_full_target_nll": 663.1445067822851
},
"branch_reference/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 154,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
154,
267,
390,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"branch_reference/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 154,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
154,
267,
390,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"operand_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 155,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
155,
281,
398,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"operand_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 155,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
155,
281,
398,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"predicate_assignment/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 351,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 36,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 60,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
8,
8,
4
],
"complete_function_curve_budgets_1_to_4": [
12,
20,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
351,
417,
478,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 20,
"summed_initial_full_target_nll": 663.1445067822849
},
"predicate_assignment/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 351,
"initial_complete_functions": 12,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 36,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 60,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
12,
8,
8,
4
],
"complete_function_curve_budgets_1_to_4": [
12,
20,
28,
32
],
"best_passing_case_curve_budgets_1_to_4": [
351,
417,
478,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 20,
"summed_initial_full_target_nll": 663.1445067822849
},
"schedule_branch/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 156,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
156,
280,
400,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"schedule_branch/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 156,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
156,
280,
400,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"schedule_operand/en": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 160,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
160,
285,
404,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
},
"schedule_operand/ko": {
"function_views": 32,
"ordered_finite_cases": 512,
"initial_cases_passed": 160,
"initial_complete_functions": 8,
"complete_functions_by_budget_four": 32,
"additional_candidate_attempts": 48,
"actual_model_predictions": 0,
"actual_feedback_predictions": 0,
"zero_call_feedback_context_declines": 0,
"type_rejections": 0,
"unattempted_masks_after_completion": 48,
"bilingual_initial_mask_disagreements": 0,
"bilingual_function_pairs": 0,
"minimum_complete_budget_histogram_1_to_4": [
8,
8,
8,
8
],
"complete_function_curve_budgets_1_to_4": [
8,
16,
24,
32
],
"best_passing_case_curve_budgets_1_to_4": [
160,
285,
404,
512
],
"summed_observed_prediction_ns": 0,
"summed_initial_probability_mass_outside_full_target": 24,
"summed_initial_full_target_nll": 663.1445067822851
}
}
}
},
"go_python_parity_max_absolute_error": {
"independent": 8.940696716308594e-8,
"joint": 1.601874828338623e-7
},
"native_calls": 0,
"new_optimizer_updates": 0,
"parity_predictions": 192,
"runtime": {
"independent-fp32": {
"warm_benchmark_predictions": 1001,
"warm_ns_per_prediction": 8932.667,
"predict_into_heap_allocations_per_call": 0,
"allocation_probe_predictions": 1001,
"resident_tensor_bytes": 50912,
"matrix_scale_bytes": 0,
"workspace_bytes": 1248
},
"independent-ptq_ternary": {
"warm_benchmark_predictions": 1001,
"warm_ns_per_prediction": 9935.583,
"predict_into_heap_allocations_per_call": 0,
"allocation_probe_predictions": 1001,
"resident_tensor_bytes": 12896,
"matrix_scale_bytes": 8,
"workspace_bytes": 1248
},
"independent-qat_ternary": {
"warm_benchmark_predictions": 1001,
"warm_ns_per_prediction": 10035.084,
"predict_into_heap_allocations_per_call": 0,
"allocation_probe_predictions": 1001,
"resident_tensor_bytes": 12896,
"matrix_scale_bytes": 8,
"workspace_bytes": 1248
},
"joint-fp32": {
"warm_benchmark_predictions": 1001,
"warm_ns_per_prediction": 10694.792,
"predict_into_heap_allocations_per_call": 0,
"allocation_probe_predictions": 1001,
"resident_tensor_bytes": 49648,
"matrix_scale_bytes": 0,
"workspace_bytes": 2160
},
"joint-ptq_ternary": {
"warm_benchmark_predictions": 1001,
"warm_ns_per_prediction": 10649,
"predict_into_heap_allocations_per_call": 0,
"allocation_probe_predictions": 1001,
"resident_tensor_bytes": 12496,
"matrix_scale_bytes": 8,
"workspace_bytes": 2160
},
"joint-qat_ternary": {
"warm_benchmark_predictions": 1001,
"warm_ns_per_prediction": 10592.958,
"predict_into_heap_allocations_per_call": 0,
"allocation_probe_predictions": 1001,
"resident_tensor_bytes": 12496,
"matrix_scale_bytes": 8,
"workspace_bytes": 2160
}
},
"schema": "gooo/joint-composition-study/v1",
"scope": "Six compositions, four goals with bilingual/parameter variants; all 16 ordered independent cases retained. Model ranks legal compiler-owned fragments and observed feedback only. Calibration selector frozen before development sessions. Four-path completeness is finite enumerative evidence, not general language understanding. Warm inference allocations exclude loading/context/session and decoded ternary matrices remain int8.",
"selected_candidate": "independent-fp32",
"source_revision": "b775ba63b81f20ec585c33a8c39f4e43b8cf5986",
"status": "PASS"
}