anhtld commited on
Commit
ee52066
·
verified ·
1 Parent(s): f82d399

Upload per-dim trainmax validation seed2

Browse files
Files changed (23) hide show
  1. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/command.txt +1 -0
  2. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/config.yaml +22 -0
  3. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/data_hash.txt +1 -0
  4. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/eval.log +0 -0
  5. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/git_hash.txt +1 -0
  6. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/measured_candidates.json +0 -0
  7. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/measured_metrics/metrics.json +0 -0
  8. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/measured_metrics/metrics_by_seed.json +1647 -0
  9. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/measured_metrics/metrics_by_task.json +182 -0
  10. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/measured_metrics/report.md +40 -0
  11. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/measured_metrics/table.tex +41 -0
  12. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/metrics.json +3041 -0
  13. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/metrics_by_seed.json +1647 -0
  14. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/metrics_by_task.json +182 -0
  15. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/report.md +13 -0
  16. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/run.log +0 -0
  17. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/source_data_hash.txt +1 -0
  18. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/source_index.json +0 -0
  19. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/source_split_hash.txt +1 -0
  20. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/split_hash.txt +1 -0
  21. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/table.tex +41 -0
  22. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/target_index.json +929 -0
  23. workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/train.log +1 -0
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/command.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ python scripts/eval_ctt_generated_rollout.py --checkpoint runs/ctt_residual_base_context_obs_seed2/model.pt --source-index data/cil_charts_rgb_refs/train/index.json --target-index data/cil_charts_rgb_refs/val/index.json --out-dir runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2 --k 8 --pool-size 0 --neighbors 8 --max-target-charts 69 --group-batch-size 1 --device cpu --sim-backend physx_cpu --render-backend cpu --restore-tolerance 1e-5 --delta-scale 1.0 --execution-action-scale 1.0 --execution-action-scale-vector 0.381401261606,0.371117178171,0.354308917074,0.237850081835,0.248271283428,0.232648812617,0.215420039658 --execution-action-transform identity --bootstrap-samples 200 --disable-action-clipping
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/config.yaml ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ bootstrap_samples: 200
2
+ checkpoint: runs/ctt_residual_base_context_obs_seed2/model.pt
3
+ delta_scale: 1.0
4
+ device: cpu
5
+ disable_action_clipping: True
6
+ exclude_self_source: False
7
+ execution_action_scale: 1.0
8
+ execution_action_scale_vector: 0.381401261606,0.371117178171,0.354308917074,0.237850081835,0.248271283428,0.232648812617,0.215420039658
9
+ execution_action_transform: identity
10
+ group_batch_size: 1
11
+ include_targets_without_positives: False
12
+ k: 8
13
+ max_target_charts: 69
14
+ neighbors: 8
15
+ out_dir: runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2
16
+ pool_size: 0
17
+ render_backend: cpu
18
+ restore_tolerance: 1e-05
19
+ sim_backend: physx_cpu
20
+ skip_metrics: False
21
+ source_index: data/cil_charts_rgb_refs/train/index.json
22
+ target_index: data/cil_charts_rgb_refs/val/index.json
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/data_hash.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ 1aacf9de98648fa36bef14e137eb43c7c6973170804650814dae41d09842c708
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/eval.log ADDED
The diff for this file is too large to render. See raw diff
 
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/git_hash.txt ADDED
@@ -0,0 +1 @@
 
 
1
+
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/measured_candidates.json ADDED
The diff for this file is too large to render. See raw diff
 
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/measured_metrics/metrics.json ADDED
The diff for this file is too large to render. See raw diff
 
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/measured_metrics/metrics_by_seed.json ADDED
@@ -0,0 +1,1647 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "1025740212": {
3
+ "any_generated_unsafe_known_at_8": 0.0,
4
+ "base_safety_label_known": 1.0,
5
+ "base_success": 0.0,
6
+ "base_unsafe_known": 0.0,
7
+ "base_utility": 0.21361251175403595,
8
+ "branch_car_at_8": 0.0009215474128723145,
9
+ "generated_safety_label_coverage_at_8": 1.0,
10
+ "generated_unsafe_rate_known_at_8": 0.0,
11
+ "hidden_chart_oracle_success_at_8": 1.0,
12
+ "hidden_chart_oracle_utility_at_8": 2.0,
13
+ "ncar_to_hidden_chart_oracle_at_8": 1.0038598175480926,
14
+ "ncar_to_proposal_oracle_at_8": 0.15426996158755663,
15
+ "outcome_ptr_at_8": 0.0,
16
+ "pairwise_causal_calibration_ece": 0.23026774639555236,
17
+ "proposal_oracle_safety_label_known_at_8": 1.0,
18
+ "proposal_oracle_success_at_8": 0.0,
19
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
20
+ "proposal_oracle_unsafe_known_at_8": 0.0,
21
+ "proposal_oracle_utility_at_8": 0.2076389193534851,
22
+ "proposal_oracle_utility_gain_over_base_at_8": -0.005973592400550842,
23
+ "selected_safety_label_known_at_8": 1.0,
24
+ "selected_success_at_8": 0.0,
25
+ "selected_success_gain_over_base_at_8": 0.0,
26
+ "selected_unsafe_known_at_8": 0.0,
27
+ "selected_utility_at_8": 0.2067173719406128,
28
+ "selected_utility_gain_over_base_at_8": -0.006895139813423157,
29
+ "selector_gap_fraction_to_hidden_at_8": 0.000515872070833396,
30
+ "selector_regret_at_8": 0.0009215474128723145,
31
+ "success_selector_gap_at_8": 0.0,
32
+ "success_support_gap_at_8": 1.0,
33
+ "success_total_car_to_hidden_at_8": 1.0,
34
+ "support_gap_at_8": 1.792361080646515,
35
+ "support_gap_fraction_to_hidden_at_8": 1.0033439510967557,
36
+ "total_car_to_hidden_at_8": 1.7932826280593872
37
+ },
38
+ "1055157354": {
39
+ "any_generated_unsafe_known_at_8": 0.0,
40
+ "base_safety_label_known": 1.0,
41
+ "base_success": 0.0,
42
+ "base_unsafe_known": 0.0,
43
+ "base_utility": 0.1879976987838745,
44
+ "branch_car_at_8": 0.014723598957061768,
45
+ "generated_safety_label_coverage_at_8": 1.0,
46
+ "generated_unsafe_rate_known_at_8": 0.0,
47
+ "hidden_chart_oracle_success_at_8": 1.0,
48
+ "hidden_chart_oracle_utility_at_8": 2.0,
49
+ "ncar_to_hidden_chart_oracle_at_8": 0.9915817866344643,
50
+ "ncar_to_proposal_oracle_at_8": 0.49115629197991395,
51
+ "outcome_ptr_at_8": 1.0,
52
+ "pairwise_causal_calibration_ece": 0.19960528924679433,
53
+ "proposal_oracle_safety_label_known_at_8": 1.0,
54
+ "proposal_oracle_success_at_8": 0.0,
55
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
56
+ "proposal_oracle_unsafe_known_at_8": 0.0,
57
+ "proposal_oracle_utility_at_8": 0.21797510981559753,
58
+ "proposal_oracle_utility_gain_over_base_at_8": 0.029977411031723022,
59
+ "selected_safety_label_known_at_8": 1.0,
60
+ "selected_success_at_8": 0.0,
61
+ "selected_success_gain_over_base_at_8": 0.0,
62
+ "selected_unsafe_known_at_8": 0.0,
63
+ "selected_utility_at_8": 0.20325151085853577,
64
+ "selected_utility_gain_over_base_at_8": 0.015253812074661255,
65
+ "selector_gap_fraction_to_hidden_at_8": 0.00812559616904461,
66
+ "selector_regret_at_8": 0.014723598957061768,
67
+ "success_selector_gap_at_8": 0.0,
68
+ "success_support_gap_at_8": 1.0,
69
+ "success_total_car_to_hidden_at_8": 1.0,
70
+ "support_gap_at_8": 1.7820248901844025,
71
+ "support_gap_fraction_to_hidden_at_8": 0.9834561959377184,
72
+ "total_car_to_hidden_at_8": 1.7967484891414642
73
+ },
74
+ "1055567618": {
75
+ "any_generated_unsafe_known_at_8": 0.0,
76
+ "base_safety_label_known": 1.0,
77
+ "base_success": 1.0,
78
+ "base_unsafe_known": 0.0,
79
+ "base_utility": 2.0,
80
+ "branch_car_at_8": 0.018837809562683105,
81
+ "generated_safety_label_coverage_at_8": 1.0,
82
+ "generated_unsafe_rate_known_at_8": 0.0,
83
+ "hidden_chart_oracle_success_at_8": 1.0,
84
+ "hidden_chart_oracle_utility_at_8": 2.0,
85
+ "ncar_to_proposal_oracle_at_8": 0.016240555308914228,
86
+ "outcome_ptr_at_8": 0.0,
87
+ "pairwise_causal_calibration_ece": 0.18844312444633177,
88
+ "proposal_oracle_safety_label_known_at_8": 1.0,
89
+ "proposal_oracle_success_at_8": 0.0,
90
+ "proposal_oracle_success_gain_over_base_at_8": -1.0,
91
+ "proposal_oracle_unsafe_known_at_8": 0.0,
92
+ "proposal_oracle_utility_at_8": 0.8400760293006897,
93
+ "proposal_oracle_utility_gain_over_base_at_8": -1.1599239706993103,
94
+ "selected_safety_label_known_at_8": 1.0,
95
+ "selected_success_at_8": 0.0,
96
+ "selected_success_gain_over_base_at_8": -1.0,
97
+ "selected_unsafe_known_at_8": 0.0,
98
+ "selected_utility_at_8": 0.8212382197380066,
99
+ "selected_utility_gain_over_base_at_8": -1.1787617802619934,
100
+ "selector_regret_at_8": 0.018837809562683105,
101
+ "success_selector_gap_at_8": 0.0,
102
+ "success_support_gap_at_8": 1.0,
103
+ "success_total_car_to_hidden_at_8": 1.0,
104
+ "support_gap_at_8": 1.1599239706993103,
105
+ "total_car_to_hidden_at_8": 1.1787617802619934
106
+ },
107
+ "1113268940": {
108
+ "any_generated_unsafe_known_at_8": 0.0,
109
+ "base_safety_label_known": 1.0,
110
+ "base_success": 0.0,
111
+ "base_unsafe_known": 0.0,
112
+ "base_utility": 0.21901588141918182,
113
+ "branch_car_at_8": 0.006647542119026184,
114
+ "generated_safety_label_coverage_at_8": 1.0,
115
+ "generated_unsafe_rate_known_at_8": 0.0,
116
+ "hidden_chart_oracle_success_at_8": 1.0,
117
+ "hidden_chart_oracle_utility_at_8": 2.0,
118
+ "ncar_to_hidden_chart_oracle_at_8": 1.0031982588625619,
119
+ "outcome_ptr_at_8": 1.0,
120
+ "pairwise_causal_calibration_ece": 0.4201805991515518,
121
+ "proposal_oracle_safety_label_known_at_8": 1.0,
122
+ "proposal_oracle_success_at_8": 0.0,
123
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
124
+ "proposal_oracle_unsafe_known_at_8": 0.0,
125
+ "proposal_oracle_utility_at_8": 0.21996736526489258,
126
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0009514838457107544,
127
+ "selected_safety_label_known_at_8": 1.0,
128
+ "selected_success_at_8": 0.0,
129
+ "selected_success_gain_over_base_at_8": 0.0,
130
+ "selected_unsafe_known_at_8": 0.0,
131
+ "selected_utility_at_8": 0.2133198231458664,
132
+ "selected_utility_gain_over_base_at_8": -0.00569605827331543,
133
+ "selector_gap_fraction_to_hidden_at_8": 0.0037325106101020683,
134
+ "selector_regret_at_8": 0.006647542119026184,
135
+ "success_selector_gap_at_8": 0.0,
136
+ "success_support_gap_at_8": 1.0,
137
+ "success_total_car_to_hidden_at_8": 1.0,
138
+ "support_gap_at_8": 1.7800326347351074,
139
+ "support_gap_fraction_to_hidden_at_8": 0.9994657538852907,
140
+ "total_car_to_hidden_at_8": 1.7866801768541336
141
+ },
142
+ "1113371261": {
143
+ "any_generated_unsafe_known_at_8": 0.0,
144
+ "base_safety_label_known": 1.0,
145
+ "base_success": 1.0,
146
+ "base_unsafe_known": 0.0,
147
+ "base_utility": 2.0,
148
+ "branch_car_at_8": 0.0,
149
+ "generated_safety_label_coverage_at_8": 1.0,
150
+ "generated_unsafe_rate_known_at_8": 0.0,
151
+ "hidden_chart_oracle_success_at_8": 1.0,
152
+ "hidden_chart_oracle_utility_at_8": 2.0,
153
+ "outcome_ptr_at_8": 0.0,
154
+ "pairwise_causal_calibration_ece": 0.10528846281262146,
155
+ "proposal_oracle_safety_label_known_at_8": 1.0,
156
+ "proposal_oracle_success_at_8": 1.0,
157
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
158
+ "proposal_oracle_unsafe_known_at_8": 0.0,
159
+ "proposal_oracle_utility_at_8": 2.0,
160
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
161
+ "selected_safety_label_known_at_8": 1.0,
162
+ "selected_success_at_8": 1.0,
163
+ "selected_success_gain_over_base_at_8": 0.0,
164
+ "selected_unsafe_known_at_8": 0.0,
165
+ "selected_utility_at_8": 2.0,
166
+ "selected_utility_gain_over_base_at_8": 0.0,
167
+ "selector_regret_at_8": 0.0,
168
+ "success_selector_gap_at_8": 0.0,
169
+ "success_support_gap_at_8": 0.0,
170
+ "success_total_car_to_hidden_at_8": 0.0,
171
+ "support_gap_at_8": 0.0,
172
+ "total_car_to_hidden_at_8": 0.0
173
+ },
174
+ "1176943004": {
175
+ "any_generated_unsafe_known_at_8": 0.0,
176
+ "base_safety_label_known": 1.0,
177
+ "base_success": 1.0,
178
+ "base_unsafe_known": 0.0,
179
+ "base_utility": 2.0,
180
+ "branch_car_at_8": 0.03303612768650055,
181
+ "generated_safety_label_coverage_at_8": 1.0,
182
+ "generated_unsafe_rate_known_at_8": 0.0,
183
+ "hidden_chart_oracle_success_at_8": 1.0,
184
+ "hidden_chart_oracle_utility_at_8": 2.0,
185
+ "ncar_to_proposal_oracle_at_8": 0.01735576759836683,
186
+ "outcome_ptr_at_8": 0.0,
187
+ "pairwise_causal_calibration_ece": 0.18726924478108561,
188
+ "proposal_oracle_safety_label_known_at_8": 1.0,
189
+ "proposal_oracle_success_at_8": 0.0,
190
+ "proposal_oracle_success_gain_over_base_at_8": -1.0,
191
+ "proposal_oracle_unsafe_known_at_8": 0.0,
192
+ "proposal_oracle_utility_at_8": 0.09653319418430328,
193
+ "proposal_oracle_utility_gain_over_base_at_8": -1.9034668058156967,
194
+ "selected_safety_label_known_at_8": 1.0,
195
+ "selected_success_at_8": 0.0,
196
+ "selected_success_gain_over_base_at_8": -1.0,
197
+ "selected_unsafe_known_at_8": 0.0,
198
+ "selected_utility_at_8": 0.06349706649780273,
199
+ "selected_utility_gain_over_base_at_8": -1.9365029335021973,
200
+ "selector_regret_at_8": 0.03303612768650055,
201
+ "success_selector_gap_at_8": 0.0,
202
+ "success_support_gap_at_8": 1.0,
203
+ "success_total_car_to_hidden_at_8": 1.0,
204
+ "support_gap_at_8": 1.9034668058156967,
205
+ "total_car_to_hidden_at_8": 1.9365029335021973
206
+ },
207
+ "1189392917": {
208
+ "any_generated_unsafe_known_at_8": 0.0,
209
+ "base_safety_label_known": 1.0,
210
+ "base_success": 1.0,
211
+ "base_unsafe_known": 0.0,
212
+ "base_utility": 2.0,
213
+ "branch_car_at_8": 0.0,
214
+ "generated_safety_label_coverage_at_8": 1.0,
215
+ "generated_unsafe_rate_known_at_8": 0.0,
216
+ "hidden_chart_oracle_success_at_8": 1.0,
217
+ "hidden_chart_oracle_utility_at_8": 2.0,
218
+ "outcome_ptr_at_8": 0.0,
219
+ "pairwise_causal_calibration_ece": 0.0,
220
+ "proposal_oracle_safety_label_known_at_8": 1.0,
221
+ "proposal_oracle_success_at_8": 1.0,
222
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
223
+ "proposal_oracle_unsafe_known_at_8": 0.0,
224
+ "proposal_oracle_utility_at_8": 2.0,
225
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
226
+ "selected_safety_label_known_at_8": 1.0,
227
+ "selected_success_at_8": 1.0,
228
+ "selected_success_gain_over_base_at_8": 0.0,
229
+ "selected_unsafe_known_at_8": 0.0,
230
+ "selected_utility_at_8": 2.0,
231
+ "selected_utility_gain_over_base_at_8": 0.0,
232
+ "selector_regret_at_8": 0.0,
233
+ "success_selector_gap_at_8": 0.0,
234
+ "success_support_gap_at_8": 0.0,
235
+ "success_total_car_to_hidden_at_8": 0.0,
236
+ "support_gap_at_8": 0.0,
237
+ "total_car_to_hidden_at_8": 0.0
238
+ },
239
+ "1231780997": {
240
+ "any_generated_unsafe_known_at_8": 0.0,
241
+ "base_safety_label_known": 1.0,
242
+ "base_success": 0.0,
243
+ "base_unsafe_known": 0.0,
244
+ "base_utility": 0.21144884824752808,
245
+ "branch_car_at_8": 0.0,
246
+ "generated_safety_label_coverage_at_8": 1.0,
247
+ "generated_unsafe_rate_known_at_8": 0.0,
248
+ "hidden_chart_oracle_success_at_8": 1.0,
249
+ "hidden_chart_oracle_utility_at_8": 2.0,
250
+ "ncar_to_hidden_chart_oracle_at_8": 1.0005967986281998,
251
+ "ncar_to_proposal_oracle_at_8": 0.0,
252
+ "outcome_ptr_at_8": 0.0,
253
+ "pairwise_causal_calibration_ece": 0.13464701141472765,
254
+ "proposal_oracle_safety_label_known_at_8": 1.0,
255
+ "proposal_oracle_success_at_8": 0.0,
256
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
257
+ "proposal_oracle_unsafe_known_at_8": 0.0,
258
+ "proposal_oracle_utility_at_8": 0.2103814333677292,
259
+ "proposal_oracle_utility_gain_over_base_at_8": -0.0010674148797988892,
260
+ "selected_safety_label_known_at_8": 1.0,
261
+ "selected_success_at_8": 0.0,
262
+ "selected_success_gain_over_base_at_8": 0.0,
263
+ "selected_unsafe_known_at_8": 0.0,
264
+ "selected_utility_at_8": 0.2103814333677292,
265
+ "selected_utility_gain_over_base_at_8": -0.0010674148797988892,
266
+ "selector_gap_fraction_to_hidden_at_8": 0.0,
267
+ "selector_regret_at_8": 0.0,
268
+ "success_selector_gap_at_8": 0.0,
269
+ "success_support_gap_at_8": 1.0,
270
+ "success_total_car_to_hidden_at_8": 1.0,
271
+ "support_gap_at_8": 1.7896185666322708,
272
+ "support_gap_fraction_to_hidden_at_8": 1.0005968042226543,
273
+ "total_car_to_hidden_at_8": 1.7896185666322708
274
+ },
275
+ "1271718007": {
276
+ "any_generated_unsafe_known_at_8": 0.0,
277
+ "base_safety_label_known": 1.0,
278
+ "base_success": 0.5,
279
+ "base_unsafe_known": 0.0,
280
+ "base_utility": 1.10479174554348,
281
+ "branch_car_at_8": 0.0,
282
+ "generated_safety_label_coverage_at_8": 1.0,
283
+ "generated_unsafe_rate_known_at_8": 0.0,
284
+ "hidden_chart_oracle_success_at_8": 1.0,
285
+ "hidden_chart_oracle_utility_at_8": 2.0,
286
+ "ncar_to_hidden_chart_oracle_at_8": 0.977550125740515,
287
+ "ncar_to_proposal_oracle_at_8": 0.0,
288
+ "outcome_ptr_at_8": 0.5,
289
+ "pairwise_causal_calibration_ece": 0.07020098827281387,
290
+ "proposal_oracle_safety_label_known_at_8": 1.0,
291
+ "proposal_oracle_success_at_8": 0.5,
292
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
293
+ "proposal_oracle_unsafe_known_at_8": 0.0,
294
+ "proposal_oracle_utility_at_8": 1.1248890534043312,
295
+ "proposal_oracle_utility_gain_over_base_at_8": 0.020097307860851288,
296
+ "selected_safety_label_known_at_8": 1.0,
297
+ "selected_success_at_8": 0.5,
298
+ "selected_success_gain_over_base_at_8": 0.0,
299
+ "selected_unsafe_known_at_8": 0.0,
300
+ "selected_utility_at_8": 1.1248890534043312,
301
+ "selected_utility_gain_over_base_at_8": 0.020097307860851288,
302
+ "selector_gap_fraction_to_hidden_at_8": 0.0,
303
+ "selector_regret_at_8": 0.0,
304
+ "success_selector_gap_at_8": 0.0,
305
+ "success_support_gap_at_8": 0.5,
306
+ "success_total_car_to_hidden_at_8": 0.5,
307
+ "support_gap_at_8": 0.8751109465956688,
308
+ "support_gap_fraction_to_hidden_at_8": 0.9775501312004184,
309
+ "total_car_to_hidden_at_8": 0.8751109465956688
310
+ },
311
+ "1273247072": {
312
+ "any_generated_unsafe_known_at_8": 0.0,
313
+ "base_safety_label_known": 1.0,
314
+ "base_success": 0.0,
315
+ "base_unsafe_known": 0.0,
316
+ "base_utility": 0.11945180594921112,
317
+ "branch_car_at_8": 0.011674322187900543,
318
+ "generated_safety_label_coverage_at_8": 1.0,
319
+ "generated_unsafe_rate_known_at_8": 0.0,
320
+ "hidden_chart_oracle_success_at_8": 1.0,
321
+ "hidden_chart_oracle_utility_at_8": 2.0,
322
+ "ncar_to_hidden_chart_oracle_at_8": 1.0157636165911068,
323
+ "ncar_to_proposal_oracle_at_8": 0.649658432039702,
324
+ "outcome_ptr_at_8": 0.0,
325
+ "pairwise_causal_calibration_ece": 0.11483608299984571,
326
+ "proposal_oracle_safety_label_known_at_8": 1.0,
327
+ "proposal_oracle_success_at_8": 0.0,
328
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
329
+ "proposal_oracle_unsafe_known_at_8": 0.0,
330
+ "proposal_oracle_utility_at_8": 0.10148187726736069,
331
+ "proposal_oracle_utility_gain_over_base_at_8": -0.017969928681850433,
332
+ "selected_safety_label_known_at_8": 1.0,
333
+ "selected_success_at_8": 0.0,
334
+ "selected_success_gain_over_base_at_8": 0.0,
335
+ "selected_unsafe_known_at_8": 0.0,
336
+ "selected_utility_at_8": 0.08980755507946014,
337
+ "selected_utility_gain_over_base_at_8": -0.029644250869750977,
338
+ "selector_gap_fraction_to_hidden_at_8": 0.006207935656652067,
339
+ "selector_regret_at_8": 0.011674322187900543,
340
+ "success_selector_gap_at_8": 0.0,
341
+ "success_support_gap_at_8": 1.0,
342
+ "success_total_car_to_hidden_at_8": 1.0,
343
+ "support_gap_at_8": 1.8985181227326393,
344
+ "support_gap_fraction_to_hidden_at_8": 1.0095556863358777,
345
+ "total_car_to_hidden_at_8": 1.9101924449205399
346
+ },
347
+ "1358421325": {
348
+ "any_generated_unsafe_known_at_8": 0.0,
349
+ "base_safety_label_known": 1.0,
350
+ "base_success": 1.0,
351
+ "base_unsafe_known": 0.0,
352
+ "base_utility": 2.0,
353
+ "branch_car_at_8": 0.6867138147354126,
354
+ "generated_safety_label_coverage_at_8": 1.0,
355
+ "generated_unsafe_rate_known_at_8": 0.0,
356
+ "hidden_chart_oracle_success_at_8": 1.0,
357
+ "hidden_chart_oracle_utility_at_8": 2.0,
358
+ "ncar_to_proposal_oracle_at_8": 0.5920648030985759,
359
+ "outcome_ptr_at_8": 0.0,
360
+ "pairwise_causal_calibration_ece": 0.24959422175257154,
361
+ "proposal_oracle_safety_label_known_at_8": 1.0,
362
+ "proposal_oracle_success_at_8": 0.0,
363
+ "proposal_oracle_success_gain_over_base_at_8": -1.0,
364
+ "proposal_oracle_unsafe_known_at_8": 0.0,
365
+ "proposal_oracle_utility_at_8": 0.8401374220848083,
366
+ "proposal_oracle_utility_gain_over_base_at_8": -1.1598625779151917,
367
+ "selected_safety_label_known_at_8": 1.0,
368
+ "selected_success_at_8": 0.0,
369
+ "selected_success_gain_over_base_at_8": -1.0,
370
+ "selected_unsafe_known_at_8": 0.0,
371
+ "selected_utility_at_8": 0.15342360734939575,
372
+ "selected_utility_gain_over_base_at_8": -1.8465763926506042,
373
+ "selector_regret_at_8": 0.6867138147354126,
374
+ "success_selector_gap_at_8": 0.0,
375
+ "success_support_gap_at_8": 1.0,
376
+ "success_total_car_to_hidden_at_8": 1.0,
377
+ "support_gap_at_8": 1.1598625779151917,
378
+ "total_car_to_hidden_at_8": 1.8465763926506042
379
+ },
380
+ "1375492786": {
381
+ "any_generated_unsafe_known_at_8": 1.0,
382
+ "base_safety_label_known": 1.0,
383
+ "base_success": 0.0,
384
+ "base_unsafe_known": 0.0,
385
+ "base_utility": 0.18323469161987305,
386
+ "branch_car_at_8": 0.0017340581398457289,
387
+ "generated_safety_label_coverage_at_8": 1.0,
388
+ "generated_unsafe_rate_known_at_8": 1.0,
389
+ "hidden_chart_oracle_success_at_8": 0.0,
390
+ "hidden_chart_oracle_utility_at_8": 0.24345096945762634,
391
+ "ncar_to_hidden_chart_oracle_at_8": 3.9862911723153553,
392
+ "ncar_to_proposal_oracle_at_8": 0.009737013067671323,
393
+ "outcome_ptr_at_8": 0.0,
394
+ "pairwise_causal_calibration_ece": 0.13157026218508128,
395
+ "proposal_oracle_safety_label_known_at_8": 1.0,
396
+ "proposal_oracle_success_at_8": 0.0,
397
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
398
+ "proposal_oracle_unsafe_known_at_8": 1.0,
399
+ "proposal_oracle_utility_at_8": 0.005145370960235596,
400
+ "proposal_oracle_utility_gain_over_base_at_8": -0.17808932065963745,
401
+ "selected_safety_label_known_at_8": 1.0,
402
+ "selected_success_at_8": 0.0,
403
+ "selected_success_gain_over_base_at_8": 0.0,
404
+ "selected_unsafe_known_at_8": 1.0,
405
+ "selected_utility_at_8": 0.003411312820389867,
406
+ "selected_utility_gain_over_base_at_8": -0.17982337879948318,
407
+ "selector_gap_fraction_to_hidden_at_8": 0.028797165851366206,
408
+ "selector_regret_at_8": 0.0017340581398457289,
409
+ "success_selector_gap_at_8": 0.0,
410
+ "success_support_gap_at_8": 0.0,
411
+ "success_total_car_to_hidden_at_8": 0.0,
412
+ "support_gap_at_8": 0.23830559849739075,
413
+ "support_gap_fraction_to_hidden_at_8": 3.957494668459602,
414
+ "total_car_to_hidden_at_8": 0.24003965663723648
415
+ },
416
+ "141984615": {
417
+ "any_generated_unsafe_known_at_8": 0.3333333333333333,
418
+ "base_safety_label_known": 1.0,
419
+ "base_success": 0.0,
420
+ "base_unsafe_known": 0.0,
421
+ "base_utility": 0.12483831991751988,
422
+ "branch_car_at_8": 0.017901106815164287,
423
+ "generated_safety_label_coverage_at_8": 1.0,
424
+ "generated_unsafe_rate_known_at_8": 0.3333333333333333,
425
+ "hidden_chart_oracle_success_at_8": 0.0,
426
+ "hidden_chart_oracle_utility_at_8": 0.20044784247875214,
427
+ "ncar_to_hidden_chart_oracle_at_8": 3.2387927981827365,
428
+ "ncar_to_proposal_oracle_at_8": 0.20565346679569363,
429
+ "outcome_ptr_at_8": 0.3333333333333333,
430
+ "pairwise_causal_calibration_ece": 0.17017158124656792,
431
+ "proposal_oracle_safety_label_known_at_8": 1.0,
432
+ "proposal_oracle_success_at_8": 0.0,
433
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
434
+ "proposal_oracle_unsafe_known_at_8": 0.3333333333333333,
435
+ "proposal_oracle_utility_at_8": 0.10611975689729054,
436
+ "proposal_oracle_utility_gain_over_base_at_8": -0.01871856302022934,
437
+ "selected_safety_label_known_at_8": 1.0,
438
+ "selected_success_at_8": 0.0,
439
+ "selected_success_gain_over_base_at_8": 0.0,
440
+ "selected_unsafe_known_at_8": 0.3333333333333333,
441
+ "selected_utility_at_8": 0.08821865008212626,
442
+ "selected_utility_gain_over_base_at_8": -0.03661966983539363,
443
+ "selector_gap_fraction_to_hidden_at_8": 0.6932408439097882,
444
+ "selector_regret_at_8": 0.017901106815164287,
445
+ "success_selector_gap_at_8": 0.0,
446
+ "success_support_gap_at_8": 0.0,
447
+ "success_total_car_to_hidden_at_8": 0.0,
448
+ "support_gap_at_8": 0.11338664591312408,
449
+ "support_gap_fraction_to_hidden_at_8": 2.951807933296765,
450
+ "total_car_to_hidden_at_8": 0.11700566027623911
451
+ },
452
+ "1475440973": {
453
+ "any_generated_unsafe_known_at_8": 0.0,
454
+ "base_safety_label_known": 1.0,
455
+ "base_success": 0.0,
456
+ "base_unsafe_known": 0.0,
457
+ "base_utility": 0.7229747772216797,
458
+ "branch_car_at_8": 0.004936397075653076,
459
+ "generated_safety_label_coverage_at_8": 1.0,
460
+ "generated_unsafe_rate_known_at_8": 0.0,
461
+ "hidden_chart_oracle_success_at_8": 1.0,
462
+ "hidden_chart_oracle_utility_at_8": 2.0,
463
+ "ncar_to_hidden_chart_oracle_at_8": 0.9997688595390668,
464
+ "ncar_to_proposal_oracle_at_8": 0.943578650052657,
465
+ "outcome_ptr_at_8": 1.0,
466
+ "pairwise_causal_calibration_ece": 0.11407560353818716,
467
+ "proposal_oracle_safety_label_known_at_8": 1.0,
468
+ "proposal_oracle_success_at_8": 0.0,
469
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
470
+ "proposal_oracle_unsafe_known_at_8": 0.0,
471
+ "proposal_oracle_utility_at_8": 0.7282063364982605,
472
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0052315592765808105,
473
+ "selected_safety_label_known_at_8": 1.0,
474
+ "selected_success_at_8": 0.0,
475
+ "selected_success_gain_over_base_at_8": 0.0,
476
+ "selected_unsafe_known_at_8": 0.0,
477
+ "selected_utility_at_8": 0.7232699394226074,
478
+ "selected_utility_gain_over_base_at_8": 0.0002951622009277344,
479
+ "selector_gap_fraction_to_hidden_at_8": 0.0038655439122129136,
480
+ "selector_regret_at_8": 0.004936397075653076,
481
+ "success_selector_gap_at_8": 0.0,
482
+ "success_support_gap_at_8": 1.0,
483
+ "success_total_car_to_hidden_at_8": 1.0,
484
+ "support_gap_at_8": 1.2717936635017395,
485
+ "support_gap_fraction_to_hidden_at_8": 0.9959033234557427,
486
+ "total_car_to_hidden_at_8": 1.2767300605773926
487
+ },
488
+ "1483645336": {
489
+ "any_generated_unsafe_known_at_8": 0.0,
490
+ "base_safety_label_known": 1.0,
491
+ "base_success": 1.0,
492
+ "base_unsafe_known": 0.0,
493
+ "base_utility": 2.0,
494
+ "branch_car_at_8": 1.8495931923389435,
495
+ "generated_safety_label_coverage_at_8": 1.0,
496
+ "generated_unsafe_rate_known_at_8": 0.0,
497
+ "hidden_chart_oracle_success_at_8": 1.0,
498
+ "hidden_chart_oracle_utility_at_8": 2.0,
499
+ "outcome_ptr_at_8": 0.0,
500
+ "pairwise_causal_calibration_ece": 0.5238458588992289,
501
+ "proposal_oracle_safety_label_known_at_8": 1.0,
502
+ "proposal_oracle_success_at_8": 1.0,
503
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
504
+ "proposal_oracle_unsafe_known_at_8": 0.0,
505
+ "proposal_oracle_utility_at_8": 2.0,
506
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
507
+ "selected_safety_label_known_at_8": 1.0,
508
+ "selected_success_at_8": 0.0,
509
+ "selected_success_gain_over_base_at_8": -1.0,
510
+ "selected_unsafe_known_at_8": 0.0,
511
+ "selected_utility_at_8": 0.15040680766105652,
512
+ "selected_utility_gain_over_base_at_8": -1.8495931923389435,
513
+ "selector_regret_at_8": 1.8495931923389435,
514
+ "success_selector_gap_at_8": 1.0,
515
+ "success_support_gap_at_8": 0.0,
516
+ "success_total_car_to_hidden_at_8": 1.0,
517
+ "support_gap_at_8": 0.0,
518
+ "total_car_to_hidden_at_8": 1.8495931923389435
519
+ },
520
+ "1522272686": {
521
+ "any_generated_unsafe_known_at_8": 0.0,
522
+ "base_safety_label_known": 1.0,
523
+ "base_success": 0.0,
524
+ "base_unsafe_known": 0.0,
525
+ "base_utility": 0.5529977083206177,
526
+ "branch_car_at_8": 0.3386564999818802,
527
+ "generated_safety_label_coverage_at_8": 1.0,
528
+ "generated_unsafe_rate_known_at_8": 0.0,
529
+ "hidden_chart_oracle_success_at_8": 1.0,
530
+ "hidden_chart_oracle_utility_at_8": 2.0,
531
+ "ncar_to_hidden_chart_oracle_at_8": 1.2270471627342485,
532
+ "ncar_to_proposal_oracle_at_8": 33.46827145305134,
533
+ "outcome_ptr_at_8": 1.0,
534
+ "pairwise_causal_calibration_ece": 0.0746614940687875,
535
+ "proposal_oracle_safety_label_known_at_8": 1.0,
536
+ "proposal_oracle_success_at_8": 0.0,
537
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
538
+ "proposal_oracle_unsafe_known_at_8": 0.0,
539
+ "proposal_oracle_utility_at_8": 0.5631164312362671,
540
+ "proposal_oracle_utility_gain_over_base_at_8": 0.010118722915649414,
541
+ "selected_safety_label_known_at_8": 1.0,
542
+ "selected_success_at_8": 0.0,
543
+ "selected_success_gain_over_base_at_8": 0.0,
544
+ "selected_unsafe_known_at_8": 0.0,
545
+ "selected_utility_at_8": 0.2244599312543869,
546
+ "selected_utility_gain_over_base_at_8": -0.3285377770662308,
547
+ "selector_gap_fraction_to_hidden_at_8": 0.23404005780034906,
548
+ "selector_regret_at_8": 0.3386564999818802,
549
+ "success_selector_gap_at_8": 0.0,
550
+ "success_support_gap_at_8": 1.0,
551
+ "success_total_car_to_hidden_at_8": 1.0,
552
+ "support_gap_at_8": 1.436883568763733,
553
+ "support_gap_fraction_to_hidden_at_8": 0.9930071134138249,
554
+ "total_car_to_hidden_at_8": 1.775540068745613
555
+ },
556
+ "1528740417": {
557
+ "any_generated_unsafe_known_at_8": 0.0,
558
+ "base_safety_label_known": 1.0,
559
+ "base_success": 0.0,
560
+ "base_unsafe_known": 0.0,
561
+ "base_utility": 0.32647061347961426,
562
+ "branch_car_at_8": 0.02146562933921814,
563
+ "generated_safety_label_coverage_at_8": 1.0,
564
+ "generated_unsafe_rate_known_at_8": 0.0,
565
+ "hidden_chart_oracle_success_at_8": 1.0,
566
+ "hidden_chart_oracle_utility_at_8": 2.0,
567
+ "ncar_to_hidden_chart_oracle_at_8": 1.0390089371449063,
568
+ "ncar_to_proposal_oracle_at_8": 0.48989279105464534,
569
+ "outcome_ptr_at_8": 0.0,
570
+ "pairwise_causal_calibration_ece": 0.13131517178411375,
571
+ "proposal_oracle_safety_label_known_at_8": 1.0,
572
+ "proposal_oracle_success_at_8": 0.0,
573
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
574
+ "proposal_oracle_unsafe_known_at_8": 0.0,
575
+ "proposal_oracle_utility_at_8": 0.2826536297798157,
576
+ "proposal_oracle_utility_gain_over_base_at_8": -0.043816983699798584,
577
+ "selected_safety_label_known_at_8": 1.0,
578
+ "selected_success_at_8": 0.0,
579
+ "selected_success_gain_over_base_at_8": 0.0,
580
+ "selected_unsafe_known_at_8": 0.0,
581
+ "selected_utility_at_8": 0.26118800044059753,
582
+ "selected_utility_gain_over_base_at_8": -0.06528261303901672,
583
+ "selector_gap_fraction_to_hidden_at_8": 0.01282656254029075,
584
+ "selector_regret_at_8": 0.02146562933921814,
585
+ "success_selector_gap_at_8": 0.0,
586
+ "success_support_gap_at_8": 1.0,
587
+ "success_total_car_to_hidden_at_8": 1.0,
588
+ "support_gap_at_8": 1.7173463702201843,
589
+ "support_gap_fraction_to_hidden_at_8": 1.026182380813105,
590
+ "total_car_to_hidden_at_8": 1.7388119995594025
591
+ },
592
+ "1549324805": {
593
+ "any_generated_unsafe_known_at_8": 1.0,
594
+ "base_safety_label_known": 1.0,
595
+ "base_success": 0.3333333333333333,
596
+ "base_unsafe_known": 0.0,
597
+ "base_utility": 0.7632389987508456,
598
+ "branch_car_at_8": 0.6191183490057787,
599
+ "generated_safety_label_coverage_at_8": 1.0,
600
+ "generated_unsafe_rate_known_at_8": 0.75,
601
+ "hidden_chart_oracle_success_at_8": 0.6666666666666666,
602
+ "hidden_chart_oracle_utility_at_8": 1.4027505218982697,
603
+ "ncar_to_hidden_chart_oracle_at_8": 2.43115209329291,
604
+ "ncar_to_proposal_oracle_at_8": 0.21532264167370982,
605
+ "outcome_ptr_at_8": 0.0,
606
+ "pairwise_causal_calibration_ece": 0.3542532249847814,
607
+ "proposal_oracle_safety_label_known_at_8": 1.0,
608
+ "proposal_oracle_success_at_8": 0.3333333333333333,
609
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
610
+ "proposal_oracle_unsafe_known_at_8": 0.6666666666666666,
611
+ "proposal_oracle_utility_at_8": 0.7099594213068485,
612
+ "proposal_oracle_utility_gain_over_base_at_8": -0.05327957744399706,
613
+ "selected_safety_label_known_at_8": 1.0,
614
+ "selected_success_at_8": 0.0,
615
+ "selected_success_gain_over_base_at_8": -0.3333333333333333,
616
+ "selected_unsafe_known_at_8": 0.3333333333333333,
617
+ "selected_utility_at_8": 0.0908410723010699,
618
+ "selected_utility_gain_over_base_at_8": -0.6723979264497757,
619
+ "selector_gap_fraction_to_hidden_at_8": 0.30888217864267864,
620
+ "selector_regret_at_8": 0.6191183490057787,
621
+ "success_selector_gap_at_8": 0.3333333333333333,
622
+ "success_support_gap_at_8": 0.3333333333333333,
623
+ "success_total_car_to_hidden_at_8": 0.6666666666666666,
624
+ "support_gap_at_8": 0.6927911005914211,
625
+ "support_gap_fraction_to_hidden_at_8": 2.1222703900724977,
626
+ "total_car_to_hidden_at_8": 1.3119094495971997
627
+ },
628
+ "1607773860": {
629
+ "any_generated_unsafe_known_at_8": 0.0,
630
+ "base_safety_label_known": 1.0,
631
+ "base_success": 0.0,
632
+ "base_unsafe_known": 0.0,
633
+ "base_utility": 0.8032726645469666,
634
+ "branch_car_at_8": 0.011432766914367676,
635
+ "generated_safety_label_coverage_at_8": 1.0,
636
+ "generated_unsafe_rate_known_at_8": 0.0,
637
+ "hidden_chart_oracle_success_at_8": 1.0,
638
+ "hidden_chart_oracle_utility_at_8": 2.0,
639
+ "ncar_to_hidden_chart_oracle_at_8": 1.0017699108028124,
640
+ "ncar_to_proposal_oracle_at_8": 1.2273941543534532,
641
+ "outcome_ptr_at_8": 1.0,
642
+ "pairwise_causal_calibration_ece": 0.251291053201395,
643
+ "proposal_oracle_safety_label_known_at_8": 1.0,
644
+ "proposal_oracle_success_at_8": 0.0,
645
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
646
+ "proposal_oracle_unsafe_known_at_8": 0.0,
647
+ "proposal_oracle_utility_at_8": 0.812587320804596,
648
+ "proposal_oracle_utility_gain_over_base_at_8": 0.009314656257629395,
649
+ "selected_safety_label_known_at_8": 1.0,
650
+ "selected_success_at_8": 0.0,
651
+ "selected_success_gain_over_base_at_8": 0.0,
652
+ "selected_unsafe_known_at_8": 0.0,
653
+ "selected_utility_at_8": 0.8011545538902283,
654
+ "selected_utility_gain_over_base_at_8": -0.0021181106567382812,
655
+ "selector_gap_fraction_to_hidden_at_8": 0.009553359880460726,
656
+ "selector_regret_at_8": 0.011432766914367676,
657
+ "success_selector_gap_at_8": 0.0,
658
+ "success_support_gap_at_8": 1.0,
659
+ "success_total_car_to_hidden_at_8": 1.0,
660
+ "support_gap_at_8": 1.187412679195404,
661
+ "support_gap_fraction_to_hidden_at_8": 0.9922165592932636,
662
+ "total_car_to_hidden_at_8": 1.1988454461097717
663
+ },
664
+ "1665739511": {
665
+ "any_generated_unsafe_known_at_8": 0.0,
666
+ "base_safety_label_known": 1.0,
667
+ "base_success": 1.0,
668
+ "base_unsafe_known": 0.0,
669
+ "base_utility": 2.0,
670
+ "branch_car_at_8": 0.0,
671
+ "generated_safety_label_coverage_at_8": 1.0,
672
+ "generated_unsafe_rate_known_at_8": 0.0,
673
+ "hidden_chart_oracle_success_at_8": 1.0,
674
+ "hidden_chart_oracle_utility_at_8": 2.0,
675
+ "outcome_ptr_at_8": 0.0,
676
+ "pairwise_causal_calibration_ece": 0.0,
677
+ "proposal_oracle_safety_label_known_at_8": 1.0,
678
+ "proposal_oracle_success_at_8": 1.0,
679
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
680
+ "proposal_oracle_unsafe_known_at_8": 0.0,
681
+ "proposal_oracle_utility_at_8": 2.0,
682
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
683
+ "selected_safety_label_known_at_8": 1.0,
684
+ "selected_success_at_8": 1.0,
685
+ "selected_success_gain_over_base_at_8": 0.0,
686
+ "selected_unsafe_known_at_8": 0.0,
687
+ "selected_utility_at_8": 2.0,
688
+ "selected_utility_gain_over_base_at_8": 0.0,
689
+ "selector_regret_at_8": 0.0,
690
+ "success_selector_gap_at_8": 0.0,
691
+ "success_support_gap_at_8": 0.0,
692
+ "success_total_car_to_hidden_at_8": 0.0,
693
+ "support_gap_at_8": 0.0,
694
+ "total_car_to_hidden_at_8": 0.0
695
+ },
696
+ "1747178030": {
697
+ "any_generated_unsafe_known_at_8": 1.0,
698
+ "base_safety_label_known": 1.0,
699
+ "base_success": 1.0,
700
+ "base_unsafe_known": 0.0,
701
+ "base_utility": 2.0,
702
+ "branch_car_at_8": 0.01012641191482544,
703
+ "generated_safety_label_coverage_at_8": 1.0,
704
+ "generated_unsafe_rate_known_at_8": 1.0,
705
+ "hidden_chart_oracle_success_at_8": 1.0,
706
+ "hidden_chart_oracle_utility_at_8": 2.0,
707
+ "ncar_to_proposal_oracle_at_8": 0.0052936433657889735,
708
+ "outcome_ptr_at_8": 0.0,
709
+ "pairwise_causal_calibration_ece": 0.16974044575049851,
710
+ "proposal_oracle_safety_label_known_at_8": 1.0,
711
+ "proposal_oracle_success_at_8": 0.0,
712
+ "proposal_oracle_success_gain_over_base_at_8": -1.0,
713
+ "proposal_oracle_unsafe_known_at_8": 1.0,
714
+ "proposal_oracle_utility_at_8": 0.08706194162368774,
715
+ "proposal_oracle_utility_gain_over_base_at_8": -1.9129380583763123,
716
+ "selected_safety_label_known_at_8": 1.0,
717
+ "selected_success_at_8": 0.0,
718
+ "selected_success_gain_over_base_at_8": -1.0,
719
+ "selected_unsafe_known_at_8": 1.0,
720
+ "selected_utility_at_8": 0.0769355297088623,
721
+ "selected_utility_gain_over_base_at_8": -1.9230644702911377,
722
+ "selector_regret_at_8": 0.01012641191482544,
723
+ "success_selector_gap_at_8": 0.0,
724
+ "success_support_gap_at_8": 1.0,
725
+ "success_total_car_to_hidden_at_8": 1.0,
726
+ "support_gap_at_8": 1.9129380583763123,
727
+ "total_car_to_hidden_at_8": 1.9230644702911377
728
+ },
729
+ "1881223807": {
730
+ "any_generated_unsafe_known_at_8": 0.0,
731
+ "base_safety_label_known": 1.0,
732
+ "base_success": 0.0,
733
+ "base_unsafe_known": 0.0,
734
+ "base_utility": 0.05987149477005005,
735
+ "branch_car_at_8": 0.0018659532070159912,
736
+ "generated_safety_label_coverage_at_8": 1.0,
737
+ "generated_unsafe_rate_known_at_8": 0.0,
738
+ "hidden_chart_oracle_success_at_8": 1.0,
739
+ "hidden_chart_oracle_utility_at_8": 2.0,
740
+ "ncar_to_hidden_chart_oracle_at_8": 1.0055300485453837,
741
+ "ncar_to_proposal_oracle_at_8": 0.21053121053631743,
742
+ "outcome_ptr_at_8": 0.0,
743
+ "pairwise_causal_calibration_ece": 0.07419200542238083,
744
+ "proposal_oracle_safety_label_known_at_8": 1.0,
745
+ "proposal_oracle_success_at_8": 0.0,
746
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
747
+ "proposal_oracle_unsafe_known_at_8": 0.0,
748
+ "proposal_oracle_utility_at_8": 0.0510084331035614,
749
+ "proposal_oracle_utility_gain_over_base_at_8": -0.008863061666488647,
750
+ "selected_safety_label_known_at_8": 1.0,
751
+ "selected_success_at_8": 0.0,
752
+ "selected_success_gain_over_base_at_8": 0.0,
753
+ "selected_unsafe_known_at_8": 0.0,
754
+ "selected_utility_at_8": 0.04914247989654541,
755
+ "selected_utility_gain_over_base_at_8": -0.010729014873504639,
756
+ "selector_gap_fraction_to_hidden_at_8": 0.0009617678426897979,
757
+ "selector_regret_at_8": 0.0018659532070159912,
758
+ "success_selector_gap_at_8": 0.0,
759
+ "success_support_gap_at_8": 1.0,
760
+ "success_total_car_to_hidden_at_8": 1.0,
761
+ "support_gap_at_8": 1.9489915668964386,
762
+ "support_gap_fraction_to_hidden_at_8": 1.0045682858854952,
763
+ "total_car_to_hidden_at_8": 1.9508575201034546
764
+ },
765
+ "1919198188": {
766
+ "any_generated_unsafe_known_at_8": 0.0,
767
+ "base_safety_label_known": 1.0,
768
+ "base_success": 0.0,
769
+ "base_unsafe_known": 0.0,
770
+ "base_utility": 0.06612768769264221,
771
+ "branch_car_at_8": 0.0067754387855529785,
772
+ "generated_safety_label_coverage_at_8": 1.0,
773
+ "generated_unsafe_rate_known_at_8": 0.0,
774
+ "hidden_chart_oracle_success_at_8": 1.0,
775
+ "hidden_chart_oracle_utility_at_8": 2.0,
776
+ "ncar_to_hidden_chart_oracle_at_8": 1.0050267795980445,
777
+ "ncar_to_proposal_oracle_at_8": 2.3000872930411496,
778
+ "outcome_ptr_at_8": 0.0,
779
+ "pairwise_causal_calibration_ece": 0.16604478136348194,
780
+ "proposal_oracle_safety_label_known_at_8": 1.0,
781
+ "proposal_oracle_success_at_8": 0.0,
782
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
783
+ "proposal_oracle_unsafe_known_at_8": 0.0,
784
+ "proposal_oracle_utility_at_8": 0.06318196654319763,
785
+ "proposal_oracle_utility_gain_over_base_at_8": -0.00294572114944458,
786
+ "selected_safety_label_known_at_8": 1.0,
787
+ "selected_success_at_8": 0.0,
788
+ "selected_success_gain_over_base_at_8": 0.0,
789
+ "selected_unsafe_known_at_8": 0.0,
790
+ "selected_utility_at_8": 0.05640652775764465,
791
+ "selected_utility_gain_over_base_at_8": -0.009721159934997559,
792
+ "selector_gap_fraction_to_hidden_at_8": 0.003503560572450107,
793
+ "selector_regret_at_8": 0.0067754387855529785,
794
+ "success_selector_gap_at_8": 0.0,
795
+ "success_support_gap_at_8": 1.0,
796
+ "success_total_car_to_hidden_at_8": 1.0,
797
+ "support_gap_at_8": 1.9368180334568024,
798
+ "support_gap_fraction_to_hidden_at_8": 1.00152322422256,
799
+ "total_car_to_hidden_at_8": 1.9435934722423553
800
+ },
801
+ "1991414747": {
802
+ "any_generated_unsafe_known_at_8": 0.0,
803
+ "base_safety_label_known": 1.0,
804
+ "base_success": 1.0,
805
+ "base_unsafe_known": 0.0,
806
+ "base_utility": 2.0,
807
+ "branch_car_at_8": 0.004616528749465942,
808
+ "generated_safety_label_coverage_at_8": 1.0,
809
+ "generated_unsafe_rate_known_at_8": 0.0,
810
+ "hidden_chart_oracle_success_at_8": 1.0,
811
+ "hidden_chart_oracle_utility_at_8": 2.0,
812
+ "ncar_to_proposal_oracle_at_8": 0.002410605043679449,
813
+ "outcome_ptr_at_8": 0.0,
814
+ "pairwise_causal_calibration_ece": 0.3416902568888008,
815
+ "proposal_oracle_safety_label_known_at_8": 1.0,
816
+ "proposal_oracle_success_at_8": 0.0,
817
+ "proposal_oracle_success_gain_over_base_at_8": -1.0,
818
+ "proposal_oracle_unsafe_known_at_8": 0.0,
819
+ "proposal_oracle_utility_at_8": 0.08490870893001556,
820
+ "proposal_oracle_utility_gain_over_base_at_8": -1.9150912910699844,
821
+ "selected_safety_label_known_at_8": 1.0,
822
+ "selected_success_at_8": 0.0,
823
+ "selected_success_gain_over_base_at_8": -1.0,
824
+ "selected_unsafe_known_at_8": 0.0,
825
+ "selected_utility_at_8": 0.08029218018054962,
826
+ "selected_utility_gain_over_base_at_8": -1.9197078198194504,
827
+ "selector_regret_at_8": 0.004616528749465942,
828
+ "success_selector_gap_at_8": 0.0,
829
+ "success_support_gap_at_8": 1.0,
830
+ "success_total_car_to_hidden_at_8": 1.0,
831
+ "support_gap_at_8": 1.9150912910699844,
832
+ "total_car_to_hidden_at_8": 1.9197078198194504
833
+ },
834
+ "1997090716": {
835
+ "any_generated_unsafe_known_at_8": 0.0,
836
+ "base_safety_label_known": 1.0,
837
+ "base_success": 1.0,
838
+ "base_unsafe_known": 0.0,
839
+ "base_utility": 2.0,
840
+ "branch_car_at_8": 0.007344692945480347,
841
+ "generated_safety_label_coverage_at_8": 1.0,
842
+ "generated_unsafe_rate_known_at_8": 0.0,
843
+ "hidden_chart_oracle_success_at_8": 1.0,
844
+ "hidden_chart_oracle_utility_at_8": 2.0,
845
+ "ncar_to_proposal_oracle_at_8": 0.003931374273492085,
846
+ "outcome_ptr_at_8": 0.0,
847
+ "pairwise_causal_calibration_ece": 0.300657409601606,
848
+ "proposal_oracle_safety_label_known_at_8": 1.0,
849
+ "proposal_oracle_success_at_8": 0.0,
850
+ "proposal_oracle_success_gain_over_base_at_8": -1.0,
851
+ "proposal_oracle_unsafe_known_at_8": 0.0,
852
+ "proposal_oracle_utility_at_8": 0.1317746937274933,
853
+ "proposal_oracle_utility_gain_over_base_at_8": -1.8682253062725067,
854
+ "selected_safety_label_known_at_8": 1.0,
855
+ "selected_success_at_8": 0.0,
856
+ "selected_success_gain_over_base_at_8": -1.0,
857
+ "selected_unsafe_known_at_8": 0.0,
858
+ "selected_utility_at_8": 0.12443000078201294,
859
+ "selected_utility_gain_over_base_at_8": -1.875569999217987,
860
+ "selector_regret_at_8": 0.007344692945480347,
861
+ "success_selector_gap_at_8": 0.0,
862
+ "success_support_gap_at_8": 1.0,
863
+ "success_total_car_to_hidden_at_8": 1.0,
864
+ "support_gap_at_8": 1.8682253062725067,
865
+ "total_car_to_hidden_at_8": 1.875569999217987
866
+ },
867
+ "2015956957": {
868
+ "any_generated_unsafe_known_at_8": 0.8,
869
+ "base_safety_label_known": 1.0,
870
+ "base_success": 0.2,
871
+ "base_unsafe_known": 0.0,
872
+ "base_utility": 0.4729121312499046,
873
+ "branch_car_at_8": 0.01736527867615223,
874
+ "generated_safety_label_coverage_at_8": 1.0,
875
+ "generated_unsafe_rate_known_at_8": 0.525,
876
+ "hidden_chart_oracle_success_at_8": 0.2,
877
+ "hidden_chart_oracle_utility_at_8": 0.5298813670873642,
878
+ "ncar_to_hidden_chart_oracle_at_8": 1.4284452600977113,
879
+ "ncar_to_proposal_oracle_at_8": 0.3825324409775967,
880
+ "outcome_ptr_at_8": 0.4,
881
+ "pairwise_causal_calibration_ece": 0.24286396292452275,
882
+ "proposal_oracle_safety_label_known_at_8": 1.0,
883
+ "proposal_oracle_success_at_8": 0.0,
884
+ "proposal_oracle_success_gain_over_base_at_8": -0.2,
885
+ "proposal_oracle_unsafe_known_at_8": 0.6,
886
+ "proposal_oracle_utility_at_8": 0.11907548382878304,
887
+ "proposal_oracle_utility_gain_over_base_at_8": -0.3538366474211216,
888
+ "selected_safety_label_known_at_8": 1.0,
889
+ "selected_success_at_8": 0.0,
890
+ "selected_success_gain_over_base_at_8": -0.2,
891
+ "selected_unsafe_known_at_8": 0.4,
892
+ "selected_utility_at_8": 0.10171020515263081,
893
+ "selected_utility_gain_over_base_at_8": -0.3712019260972738,
894
+ "selector_gap_fraction_to_hidden_at_8": 0.24547848203994205,
895
+ "selector_regret_at_8": 0.01736527867615223,
896
+ "success_selector_gap_at_8": 0.0,
897
+ "success_support_gap_at_8": 0.2,
898
+ "success_total_car_to_hidden_at_8": 0.2,
899
+ "support_gap_at_8": 0.4130037985742092,
900
+ "support_gap_fraction_to_hidden_at_8": 1.2234746297648738,
901
+ "total_car_to_hidden_at_8": 0.42817116193473337
902
+ },
903
+ "2054899126": {
904
+ "any_generated_unsafe_known_at_8": 1.0,
905
+ "base_safety_label_known": 1.0,
906
+ "base_success": 1.0,
907
+ "base_unsafe_known": 0.0,
908
+ "base_utility": 2.0,
909
+ "branch_car_at_8": 0.0,
910
+ "generated_safety_label_coverage_at_8": 1.0,
911
+ "generated_unsafe_rate_known_at_8": 1.0,
912
+ "hidden_chart_oracle_success_at_8": 1.0,
913
+ "hidden_chart_oracle_utility_at_8": 2.0,
914
+ "outcome_ptr_at_8": 0.0,
915
+ "pairwise_causal_calibration_ece": 0.0,
916
+ "proposal_oracle_safety_label_known_at_8": 1.0,
917
+ "proposal_oracle_success_at_8": 1.0,
918
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
919
+ "proposal_oracle_unsafe_known_at_8": 1.0,
920
+ "proposal_oracle_utility_at_8": 2.0,
921
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
922
+ "selected_safety_label_known_at_8": 1.0,
923
+ "selected_success_at_8": 1.0,
924
+ "selected_success_gain_over_base_at_8": 0.0,
925
+ "selected_unsafe_known_at_8": 1.0,
926
+ "selected_utility_at_8": 2.0,
927
+ "selected_utility_gain_over_base_at_8": 0.0,
928
+ "selector_regret_at_8": 0.0,
929
+ "success_selector_gap_at_8": 0.0,
930
+ "success_support_gap_at_8": 0.0,
931
+ "success_total_car_to_hidden_at_8": 0.0,
932
+ "support_gap_at_8": 0.0,
933
+ "total_car_to_hidden_at_8": 0.0
934
+ },
935
+ "211245842": {
936
+ "any_generated_unsafe_known_at_8": 0.0,
937
+ "base_safety_label_known": 1.0,
938
+ "base_success": 0.0,
939
+ "base_unsafe_known": 0.0,
940
+ "base_utility": 0.791279137134552,
941
+ "branch_car_at_8": 0.04498988389968872,
942
+ "generated_safety_label_coverage_at_8": 1.0,
943
+ "generated_unsafe_rate_known_at_8": 0.0,
944
+ "hidden_chart_oracle_success_at_8": 1.0,
945
+ "hidden_chart_oracle_utility_at_8": 2.0,
946
+ "ncar_to_hidden_chart_oracle_at_8": 1.0007633933743367,
947
+ "ncar_to_proposal_oracle_at_8": 1.0209391670158559,
948
+ "outcome_ptr_at_8": 1.0,
949
+ "pairwise_causal_calibration_ece": 0.4141064151265617,
950
+ "proposal_oracle_safety_label_known_at_8": 1.0,
951
+ "proposal_oracle_success_at_8": 0.0,
952
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
953
+ "proposal_oracle_unsafe_known_at_8": 0.0,
954
+ "proposal_oracle_utility_at_8": 0.8353462815284729,
955
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0440671443939209,
956
+ "selected_safety_label_known_at_8": 1.0,
957
+ "selected_success_at_8": 0.0,
958
+ "selected_success_gain_over_base_at_8": 0.0,
959
+ "selected_unsafe_known_at_8": 0.0,
960
+ "selected_utility_at_8": 0.7903563976287842,
961
+ "selected_utility_gain_over_base_at_8": -0.0009227395057678223,
962
+ "selector_gap_fraction_to_hidden_at_8": 0.0372210700434455,
963
+ "selector_regret_at_8": 0.04498988389968872,
964
+ "success_selector_gap_at_8": 0.0,
965
+ "success_support_gap_at_8": 1.0,
966
+ "success_total_car_to_hidden_at_8": 1.0,
967
+ "support_gap_at_8": 1.164653718471527,
968
+ "support_gap_fraction_to_hidden_at_8": 0.9635423316104156,
969
+ "total_car_to_hidden_at_8": 1.2096436023712158
970
+ },
971
+ "2141802816": {
972
+ "any_generated_unsafe_known_at_8": 0.0,
973
+ "base_safety_label_known": 1.0,
974
+ "base_success": 0.0,
975
+ "base_unsafe_known": 0.0,
976
+ "base_utility": 0.23664958775043488,
977
+ "branch_car_at_8": 0.038818299770355225,
978
+ "generated_safety_label_coverage_at_8": 1.0,
979
+ "generated_unsafe_rate_known_at_8": 0.0,
980
+ "hidden_chart_oracle_success_at_8": 1.0,
981
+ "hidden_chart_oracle_utility_at_8": 2.0,
982
+ "ncar_to_hidden_chart_oracle_at_8": 1.0465578663511705,
983
+ "ncar_to_proposal_oracle_at_8": 0.8969200607232635,
984
+ "outcome_ptr_at_8": 0.0,
985
+ "pairwise_causal_calibration_ece": 0.3985289040635325,
986
+ "proposal_oracle_safety_label_known_at_8": 1.0,
987
+ "proposal_oracle_success_at_8": 0.0,
988
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
989
+ "proposal_oracle_unsafe_known_at_8": 0.0,
990
+ "proposal_oracle_utility_at_8": 0.1933700442314148,
991
+ "proposal_oracle_utility_gain_over_base_at_8": -0.04327954351902008,
992
+ "selected_safety_label_known_at_8": 1.0,
993
+ "selected_success_at_8": 0.0,
994
+ "selected_success_gain_over_base_at_8": 0.0,
995
+ "selected_unsafe_known_at_8": 0.0,
996
+ "selected_utility_at_8": 0.15455174446105957,
997
+ "selected_utility_gain_over_base_at_8": -0.0820978432893753,
998
+ "selector_gap_fraction_to_hidden_at_8": 0.022013945441980202,
999
+ "selector_regret_at_8": 0.038818299770355225,
1000
+ "success_selector_gap_at_8": 0.0,
1001
+ "success_support_gap_at_8": 1.0,
1002
+ "success_total_car_to_hidden_at_8": 1.0,
1003
+ "support_gap_at_8": 1.8066299557685852,
1004
+ "support_gap_fraction_to_hidden_at_8": 1.0245439268442436,
1005
+ "total_car_to_hidden_at_8": 1.8454482555389404
1006
+ },
1007
+ "256770874": {
1008
+ "any_generated_unsafe_known_at_8": 0.0,
1009
+ "base_safety_label_known": 1.0,
1010
+ "base_success": 0.0,
1011
+ "base_unsafe_known": 0.0,
1012
+ "base_utility": 0.18160098791122437,
1013
+ "branch_car_at_8": 0.042443424463272095,
1014
+ "generated_safety_label_coverage_at_8": 1.0,
1015
+ "generated_unsafe_rate_known_at_8": 0.0,
1016
+ "hidden_chart_oracle_success_at_8": 1.0,
1017
+ "hidden_chart_oracle_utility_at_8": 2.0,
1018
+ "ncar_to_hidden_chart_oracle_at_8": 1.002821760012782,
1019
+ "ncar_to_proposal_oracle_at_8": 1.1375171273285567,
1020
+ "outcome_ptr_at_8": 1.0,
1021
+ "pairwise_causal_calibration_ece": 0.26681470238264204,
1022
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1023
+ "proposal_oracle_success_at_8": 0.0,
1024
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1025
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1026
+ "proposal_oracle_utility_at_8": 0.21891331672668457,
1027
+ "proposal_oracle_utility_gain_over_base_at_8": 0.037312328815460205,
1028
+ "selected_safety_label_known_at_8": 1.0,
1029
+ "selected_success_at_8": 0.0,
1030
+ "selected_success_gain_over_base_at_8": 0.0,
1031
+ "selected_unsafe_known_at_8": 0.0,
1032
+ "selected_utility_at_8": 0.17646989226341248,
1033
+ "selected_utility_gain_over_base_at_8": -0.00513109564781189,
1034
+ "selector_gap_fraction_to_hidden_at_8": 0.023341095205786425,
1035
+ "selector_regret_at_8": 0.042443424463272095,
1036
+ "success_selector_gap_at_8": 0.0,
1037
+ "success_support_gap_at_8": 1.0,
1038
+ "success_total_car_to_hidden_at_8": 1.0,
1039
+ "support_gap_at_8": 1.7810866832733154,
1040
+ "support_gap_fraction_to_hidden_at_8": 0.9794806703218564,
1041
+ "total_car_to_hidden_at_8": 1.8235301077365875
1042
+ },
1043
+ "278008524": {
1044
+ "any_generated_unsafe_known_at_8": 0.5,
1045
+ "base_safety_label_known": 1.0,
1046
+ "base_success": 0.0,
1047
+ "base_unsafe_known": 0.0,
1048
+ "base_utility": 0.060655687004327774,
1049
+ "branch_car_at_8": 0.008101326413452625,
1050
+ "generated_safety_label_coverage_at_8": 1.0,
1051
+ "generated_unsafe_rate_known_at_8": 0.5,
1052
+ "hidden_chart_oracle_success_at_8": 0.5,
1053
+ "hidden_chart_oracle_utility_at_8": 1.113811194896698,
1054
+ "ncar_to_hidden_chart_oracle_at_8": 1.1481789711765635,
1055
+ "ncar_to_proposal_oracle_at_8": 0.4289430850076562,
1056
+ "outcome_ptr_at_8": 0.0,
1057
+ "pairwise_causal_calibration_ece": 0.20536892413092878,
1058
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1059
+ "proposal_oracle_success_at_8": 0.0,
1060
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1061
+ "proposal_oracle_unsafe_known_at_8": 0.5,
1062
+ "proposal_oracle_utility_at_8": 0.03437099326401949,
1063
+ "proposal_oracle_utility_gain_over_base_at_8": -0.026284693740308285,
1064
+ "selected_safety_label_known_at_8": 1.0,
1065
+ "selected_success_at_8": 0.0,
1066
+ "selected_success_gain_over_base_at_8": 0.0,
1067
+ "selected_unsafe_known_at_8": 0.5,
1068
+ "selected_utility_at_8": 0.026269666850566864,
1069
+ "selected_utility_gain_over_base_at_8": -0.03438602015376091,
1070
+ "selector_gap_fraction_to_hidden_at_8": 0.024682996062310017,
1071
+ "selector_regret_at_8": 0.008101326413452625,
1072
+ "success_selector_gap_at_8": 0.0,
1073
+ "success_support_gap_at_8": 0.5,
1074
+ "success_total_car_to_hidden_at_8": 0.5,
1075
+ "support_gap_at_8": 1.0794402016326785,
1076
+ "support_gap_fraction_to_hidden_at_8": 1.1234960169117114,
1077
+ "total_car_to_hidden_at_8": 1.0875415280461311
1078
+ },
1079
+ "416183106": {
1080
+ "any_generated_unsafe_known_at_8": 0.0,
1081
+ "base_safety_label_known": 1.0,
1082
+ "base_success": 0.0,
1083
+ "base_unsafe_known": 0.0,
1084
+ "base_utility": 0.23001998662948608,
1085
+ "branch_car_at_8": 0.0,
1086
+ "generated_safety_label_coverage_at_8": 1.0,
1087
+ "generated_unsafe_rate_known_at_8": 0.0,
1088
+ "hidden_chart_oracle_success_at_8": 1.0,
1089
+ "hidden_chart_oracle_utility_at_8": 2.0,
1090
+ "ncar_to_hidden_chart_oracle_at_8": 0.828016730744017,
1091
+ "ncar_to_proposal_oracle_at_8": 0.0,
1092
+ "outcome_ptr_at_8": 1.0,
1093
+ "pairwise_causal_calibration_ece": 0.07874347996488716,
1094
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1095
+ "proposal_oracle_success_at_8": 0.0,
1096
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1097
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1098
+ "proposal_oracle_utility_at_8": 0.5344269275665283,
1099
+ "proposal_oracle_utility_gain_over_base_at_8": 0.30440694093704224,
1100
+ "selected_safety_label_known_at_8": 1.0,
1101
+ "selected_success_at_8": 0.0,
1102
+ "selected_success_gain_over_base_at_8": 0.0,
1103
+ "selected_unsafe_known_at_8": 0.0,
1104
+ "selected_utility_at_8": 0.5344269275665283,
1105
+ "selected_utility_gain_over_base_at_8": 0.30440694093704224,
1106
+ "selector_gap_fraction_to_hidden_at_8": 0.0,
1107
+ "selector_regret_at_8": 0.0,
1108
+ "success_selector_gap_at_8": 0.0,
1109
+ "success_support_gap_at_8": 1.0,
1110
+ "success_total_car_to_hidden_at_8": 1.0,
1111
+ "support_gap_at_8": 1.4655730724334717,
1112
+ "support_gap_fraction_to_hidden_at_8": 0.8280167354221304,
1113
+ "total_car_to_hidden_at_8": 1.4655730724334717
1114
+ },
1115
+ "534717141": {
1116
+ "any_generated_unsafe_known_at_8": 0.0,
1117
+ "base_safety_label_known": 1.0,
1118
+ "base_success": 0.0,
1119
+ "base_unsafe_known": 0.0,
1120
+ "base_utility": 0.13276933133602142,
1121
+ "branch_car_at_8": 0.0,
1122
+ "generated_safety_label_coverage_at_8": 1.0,
1123
+ "generated_unsafe_rate_known_at_8": 0.0,
1124
+ "hidden_chart_oracle_success_at_8": 1.0,
1125
+ "hidden_chart_oracle_utility_at_8": 2.0,
1126
+ "ncar_to_hidden_chart_oracle_at_8": 1.0230647928623209,
1127
+ "ncar_to_proposal_oracle_at_8": 0.0,
1128
+ "outcome_ptr_at_8": 0.0,
1129
+ "pairwise_causal_calibration_ece": 0.0753952117725891,
1130
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1131
+ "proposal_oracle_success_at_8": 0.0,
1132
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1133
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1134
+ "proposal_oracle_utility_at_8": 0.08970203250646591,
1135
+ "proposal_oracle_utility_gain_over_base_at_8": -0.04306729882955551,
1136
+ "selected_safety_label_known_at_8": 1.0,
1137
+ "selected_success_at_8": 0.0,
1138
+ "selected_success_gain_over_base_at_8": 0.0,
1139
+ "selected_unsafe_known_at_8": 0.0,
1140
+ "selected_utility_at_8": 0.08970203250646591,
1141
+ "selected_utility_gain_over_base_at_8": -0.04306729882955551,
1142
+ "selector_gap_fraction_to_hidden_at_8": 0.0,
1143
+ "selector_regret_at_8": 0.0,
1144
+ "success_selector_gap_at_8": 0.0,
1145
+ "success_support_gap_at_8": 1.0,
1146
+ "success_total_car_to_hidden_at_8": 1.0,
1147
+ "support_gap_at_8": 1.910297967493534,
1148
+ "support_gap_fraction_to_hidden_at_8": 1.0230647983413697,
1149
+ "total_car_to_hidden_at_8": 1.910297967493534
1150
+ },
1151
+ "551504635": {
1152
+ "any_generated_unsafe_known_at_8": 0.0,
1153
+ "base_safety_label_known": 1.0,
1154
+ "base_success": 0.0,
1155
+ "base_unsafe_known": 0.0,
1156
+ "base_utility": 0.7272467017173767,
1157
+ "branch_car_at_8": 1.2730597853660583,
1158
+ "generated_safety_label_coverage_at_8": 1.0,
1159
+ "generated_unsafe_rate_known_at_8": 0.0,
1160
+ "hidden_chart_oracle_success_at_8": 1.0,
1161
+ "hidden_chart_oracle_utility_at_8": 2.0,
1162
+ "ncar_to_hidden_chart_oracle_at_8": 1.0002407984968027,
1163
+ "ncar_to_proposal_oracle_at_8": 1.0002407984968027,
1164
+ "outcome_ptr_at_8": 1.0,
1165
+ "pairwise_causal_calibration_ece": 0.2344395807565722,
1166
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1167
+ "proposal_oracle_success_at_8": 1.0,
1168
+ "proposal_oracle_success_gain_over_base_at_8": 1.0,
1169
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1170
+ "proposal_oracle_utility_at_8": 2.0,
1171
+ "proposal_oracle_utility_gain_over_base_at_8": 1.2727532982826233,
1172
+ "selected_safety_label_known_at_8": 1.0,
1173
+ "selected_success_at_8": 0.0,
1174
+ "selected_success_gain_over_base_at_8": 0.0,
1175
+ "selected_unsafe_known_at_8": 0.0,
1176
+ "selected_utility_at_8": 0.7269402146339417,
1177
+ "selected_utility_gain_over_base_at_8": -0.0003064870834350586,
1178
+ "selector_gap_fraction_to_hidden_at_8": 1.0002408063556767,
1179
+ "selector_regret_at_8": 1.2730597853660583,
1180
+ "success_selector_gap_at_8": 1.0,
1181
+ "success_support_gap_at_8": 0.0,
1182
+ "success_total_car_to_hidden_at_8": 1.0,
1183
+ "support_gap_at_8": 0.0,
1184
+ "support_gap_fraction_to_hidden_at_8": 0.0,
1185
+ "total_car_to_hidden_at_8": 1.2730597853660583
1186
+ },
1187
+ "565369392": {
1188
+ "any_generated_unsafe_known_at_8": 1.0,
1189
+ "base_safety_label_known": 1.0,
1190
+ "base_success": 0.3333333333333333,
1191
+ "base_unsafe_known": 0.0,
1192
+ "base_utility": 0.8445292512575785,
1193
+ "branch_car_at_8": 0.0,
1194
+ "generated_safety_label_coverage_at_8": 1.0,
1195
+ "generated_unsafe_rate_known_at_8": 0.8333333333333334,
1196
+ "hidden_chart_oracle_success_at_8": 0.6666666666666666,
1197
+ "hidden_chart_oracle_utility_at_8": 1.4266315797964733,
1198
+ "ncar_to_hidden_chart_oracle_at_8": 4.740664659244946,
1199
+ "ncar_to_proposal_oracle_at_8": 0.0,
1200
+ "outcome_ptr_at_8": 0.3333333333333333,
1201
+ "pairwise_causal_calibration_ece": 0.13140272637925274,
1202
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1203
+ "proposal_oracle_success_at_8": 0.0,
1204
+ "proposal_oracle_success_gain_over_base_at_8": -0.3333333333333333,
1205
+ "proposal_oracle_unsafe_known_at_8": 0.6666666666666666,
1206
+ "proposal_oracle_utility_at_8": 0.20857956757148108,
1207
+ "proposal_oracle_utility_gain_over_base_at_8": -0.6359496836860975,
1208
+ "selected_safety_label_known_at_8": 1.0,
1209
+ "selected_success_at_8": 0.0,
1210
+ "selected_success_gain_over_base_at_8": -0.3333333333333333,
1211
+ "selected_unsafe_known_at_8": 0.6666666666666666,
1212
+ "selected_utility_at_8": 0.20857956757148108,
1213
+ "selected_utility_gain_over_base_at_8": -0.6359496836860975,
1214
+ "selector_gap_fraction_to_hidden_at_8": 0.0,
1215
+ "selector_regret_at_8": 0.0,
1216
+ "success_selector_gap_at_8": 0.0,
1217
+ "success_support_gap_at_8": 0.6666666666666666,
1218
+ "success_total_car_to_hidden_at_8": 0.6666666666666666,
1219
+ "support_gap_at_8": 1.218052012224992,
1220
+ "support_gap_fraction_to_hidden_at_8": 4.740673332898252,
1221
+ "total_car_to_hidden_at_8": 1.218052012224992
1222
+ },
1223
+ "565621826": {
1224
+ "any_generated_unsafe_known_at_8": 1.0,
1225
+ "base_safety_label_known": 1.0,
1226
+ "base_success": 1.0,
1227
+ "base_unsafe_known": 0.0,
1228
+ "base_utility": 2.0,
1229
+ "branch_car_at_8": 0.0,
1230
+ "generated_safety_label_coverage_at_8": 1.0,
1231
+ "generated_unsafe_rate_known_at_8": 1.0,
1232
+ "hidden_chart_oracle_success_at_8": 1.0,
1233
+ "hidden_chart_oracle_utility_at_8": 2.0,
1234
+ "outcome_ptr_at_8": 0.0,
1235
+ "pairwise_causal_calibration_ece": 0.0,
1236
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1237
+ "proposal_oracle_success_at_8": 1.0,
1238
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1239
+ "proposal_oracle_unsafe_known_at_8": 1.0,
1240
+ "proposal_oracle_utility_at_8": 2.0,
1241
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
1242
+ "selected_safety_label_known_at_8": 1.0,
1243
+ "selected_success_at_8": 1.0,
1244
+ "selected_success_gain_over_base_at_8": 0.0,
1245
+ "selected_unsafe_known_at_8": 1.0,
1246
+ "selected_utility_at_8": 2.0,
1247
+ "selected_utility_gain_over_base_at_8": 0.0,
1248
+ "selector_regret_at_8": 0.0,
1249
+ "success_selector_gap_at_8": 0.0,
1250
+ "success_support_gap_at_8": 0.0,
1251
+ "success_total_car_to_hidden_at_8": 0.0,
1252
+ "support_gap_at_8": 0.0,
1253
+ "total_car_to_hidden_at_8": 0.0
1254
+ },
1255
+ "585904130": {
1256
+ "any_generated_unsafe_known_at_8": 0.3333333333333333,
1257
+ "base_safety_label_known": 1.0,
1258
+ "base_success": 0.0,
1259
+ "base_unsafe_known": 0.0,
1260
+ "base_utility": 0.13642114897569022,
1261
+ "branch_car_at_8": 0.022781788061062496,
1262
+ "generated_safety_label_coverage_at_8": 1.0,
1263
+ "generated_unsafe_rate_known_at_8": 0.3333333333333333,
1264
+ "hidden_chart_oracle_success_at_8": 0.0,
1265
+ "hidden_chart_oracle_utility_at_8": 0.23764699200789133,
1266
+ "ncar_to_hidden_chart_oracle_at_8": 1.5916216907065317,
1267
+ "ncar_to_proposal_oracle_at_8": 0.5193742667883882,
1268
+ "outcome_ptr_at_8": 0.3333333333333333,
1269
+ "pairwise_causal_calibration_ece": 0.17283993917588428,
1270
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1271
+ "proposal_oracle_success_at_8": 0.0,
1272
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1273
+ "proposal_oracle_unsafe_known_at_8": 0.3333333333333333,
1274
+ "proposal_oracle_utility_at_8": 0.09142490103840828,
1275
+ "proposal_oracle_utility_gain_over_base_at_8": -0.04499624793728193,
1276
+ "selected_safety_label_known_at_8": 1.0,
1277
+ "selected_success_at_8": 0.0,
1278
+ "selected_success_gain_over_base_at_8": 0.0,
1279
+ "selected_unsafe_known_at_8": 0.3333333333333333,
1280
+ "selected_utility_at_8": 0.06864311297734578,
1281
+ "selected_utility_gain_over_base_at_8": -0.06777803599834442,
1282
+ "selector_gap_fraction_to_hidden_at_8": 0.37677313988714195,
1283
+ "selector_regret_at_8": 0.022781788061062496,
1284
+ "success_selector_gap_at_8": 0.0,
1285
+ "success_support_gap_at_8": 0.0,
1286
+ "success_total_car_to_hidden_at_8": 0.0,
1287
+ "support_gap_at_8": 0.14622209096948305,
1288
+ "support_gap_fraction_to_hidden_at_8": 1.2148487255248963,
1289
+ "total_car_to_hidden_at_8": 0.16900387903054556
1290
+ },
1291
+ "626486813": {
1292
+ "any_generated_unsafe_known_at_8": 0.5,
1293
+ "base_safety_label_known": 1.0,
1294
+ "base_success": 0.25,
1295
+ "base_unsafe_known": 0.0,
1296
+ "base_utility": 0.5452363695949316,
1297
+ "branch_car_at_8": 0.01009321678429842,
1298
+ "generated_safety_label_coverage_at_8": 1.0,
1299
+ "generated_unsafe_rate_known_at_8": 0.28125,
1300
+ "hidden_chart_oracle_success_at_8": 0.25,
1301
+ "hidden_chart_oracle_utility_at_8": 0.6124684102833271,
1302
+ "ncar_to_hidden_chart_oracle_at_8": 0.8102411086604181,
1303
+ "ncar_to_proposal_oracle_at_8": 0.2258722711825077,
1304
+ "outcome_ptr_at_8": 0.25,
1305
+ "pairwise_causal_calibration_ece": 0.2769376371795747,
1306
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1307
+ "proposal_oracle_success_at_8": 0.0,
1308
+ "proposal_oracle_success_gain_over_base_at_8": -0.25,
1309
+ "proposal_oracle_unsafe_known_at_8": 0.25,
1310
+ "proposal_oracle_utility_at_8": 0.09146204218268394,
1311
+ "proposal_oracle_utility_gain_over_base_at_8": -0.45377432741224766,
1312
+ "selected_safety_label_known_at_8": 1.0,
1313
+ "selected_success_at_8": 0.0,
1314
+ "selected_success_gain_over_base_at_8": -0.25,
1315
+ "selected_unsafe_known_at_8": 0.25,
1316
+ "selected_utility_at_8": 0.08136882539838552,
1317
+ "selected_utility_gain_over_base_at_8": -0.4638675441965461,
1318
+ "selector_gap_fraction_to_hidden_at_8": 0.3592836940693531,
1319
+ "selector_regret_at_8": 0.01009321678429842,
1320
+ "success_selector_gap_at_8": 0.0,
1321
+ "success_support_gap_at_8": 0.25,
1322
+ "success_total_car_to_hidden_at_8": 0.25,
1323
+ "support_gap_at_8": 0.5353103466331959,
1324
+ "support_gap_fraction_to_hidden_at_8": 0.7858811145047201,
1325
+ "total_car_to_hidden_at_8": 0.5373998517170548
1326
+ },
1327
+ "636680137": {
1328
+ "any_generated_unsafe_known_at_8": 1.0,
1329
+ "base_safety_label_known": 1.0,
1330
+ "base_success": 0.0,
1331
+ "base_unsafe_known": 0.0,
1332
+ "base_utility": 0.07024133205413818,
1333
+ "branch_car_at_8": 0.02178390882909298,
1334
+ "generated_safety_label_coverage_at_8": 1.0,
1335
+ "generated_unsafe_rate_known_at_8": 0.25,
1336
+ "hidden_chart_oracle_success_at_8": 1.0,
1337
+ "hidden_chart_oracle_utility_at_8": 2.0,
1338
+ "ncar_to_hidden_chart_oracle_at_8": 1.0226806271202884,
1339
+ "ncar_to_proposal_oracle_at_8": 0.9908871412464478,
1340
+ "outcome_ptr_at_8": 0.0,
1341
+ "pairwise_causal_calibration_ece": 0.5932483433892377,
1342
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1343
+ "proposal_oracle_success_at_8": 0.0,
1344
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1345
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1346
+ "proposal_oracle_utility_at_8": 0.048257093876600266,
1347
+ "proposal_oracle_utility_gain_over_base_at_8": -0.021984238177537918,
1348
+ "selected_safety_label_known_at_8": 1.0,
1349
+ "selected_success_at_8": 0.0,
1350
+ "selected_success_gain_over_base_at_8": 0.0,
1351
+ "selected_unsafe_known_at_8": 0.0,
1352
+ "selected_utility_at_8": 0.026473185047507286,
1353
+ "selected_utility_gain_over_base_at_8": -0.0437681470066309,
1354
+ "selector_gap_fraction_to_hidden_at_8": 0.011288410924605892,
1355
+ "selector_regret_at_8": 0.02178390882909298,
1356
+ "success_selector_gap_at_8": 0.0,
1357
+ "success_support_gap_at_8": 1.0,
1358
+ "success_total_car_to_hidden_at_8": 1.0,
1359
+ "support_gap_at_8": 1.9517429061233997,
1360
+ "support_gap_fraction_to_hidden_at_8": 1.0113922214952087,
1361
+ "total_car_to_hidden_at_8": 1.9735268149524927
1362
+ },
1363
+ "706479600": {
1364
+ "any_generated_unsafe_known_at_8": 0.0,
1365
+ "base_safety_label_known": 1.0,
1366
+ "base_success": 0.0,
1367
+ "base_unsafe_known": 0.0,
1368
+ "base_utility": 0.8188905119895935,
1369
+ "branch_car_at_8": 0.018285930156707764,
1370
+ "generated_safety_label_coverage_at_8": 1.0,
1371
+ "generated_unsafe_rate_known_at_8": 0.0,
1372
+ "hidden_chart_oracle_success_at_8": 1.0,
1373
+ "hidden_chart_oracle_utility_at_8": 2.0,
1374
+ "ncar_to_hidden_chart_oracle_at_8": 1.0001120742118061,
1375
+ "ncar_to_proposal_oracle_at_8": 1.0072917889865605,
1376
+ "outcome_ptr_at_8": 1.0,
1377
+ "pairwise_causal_calibration_ece": 0.28238485246274814,
1378
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1379
+ "proposal_oracle_success_at_8": 0.0,
1380
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1381
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1382
+ "proposal_oracle_utility_at_8": 0.8370440602302551,
1383
+ "proposal_oracle_utility_gain_over_base_at_8": 0.01815354824066162,
1384
+ "selected_safety_label_known_at_8": 1.0,
1385
+ "selected_success_at_8": 0.0,
1386
+ "selected_success_gain_over_base_at_8": 0.0,
1387
+ "selected_unsafe_known_at_8": 0.0,
1388
+ "selected_utility_at_8": 0.8187581300735474,
1389
+ "selected_utility_gain_over_base_at_8": -0.00013238191604614258,
1390
+ "selector_gap_fraction_to_hidden_at_8": 0.015481994126988717,
1391
+ "selector_regret_at_8": 0.018285930156707764,
1392
+ "success_selector_gap_at_8": 0.0,
1393
+ "success_support_gap_at_8": 1.0,
1394
+ "success_total_car_to_hidden_at_8": 1.0,
1395
+ "support_gap_at_8": 1.1629559397697449,
1396
+ "support_gap_fraction_to_hidden_at_8": 0.9846300885523818,
1397
+ "total_car_to_hidden_at_8": 1.1812418699264526
1398
+ },
1399
+ "750448296": {
1400
+ "any_generated_unsafe_known_at_8": 0.5,
1401
+ "base_safety_label_known": 1.0,
1402
+ "base_success": 0.0,
1403
+ "base_unsafe_known": 0.0,
1404
+ "base_utility": 0.14706942811608315,
1405
+ "branch_car_at_8": 0.02600152976810932,
1406
+ "generated_safety_label_coverage_at_8": 1.0,
1407
+ "generated_unsafe_rate_known_at_8": 0.1875,
1408
+ "hidden_chart_oracle_success_at_8": 0.0,
1409
+ "hidden_chart_oracle_utility_at_8": 0.23580759018659592,
1410
+ "ncar_to_hidden_chart_oracle_at_8": 4.978844987783829,
1411
+ "ncar_to_proposal_oracle_at_8": 5.576871543359603,
1412
+ "outcome_ptr_at_8": 0.5,
1413
+ "pairwise_causal_calibration_ece": 0.19889074905685378,
1414
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1415
+ "proposal_oracle_success_at_8": 0.0,
1416
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1417
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1418
+ "proposal_oracle_utility_at_8": 0.14822042733430862,
1419
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0011509992182254791,
1420
+ "selected_safety_label_known_at_8": 1.0,
1421
+ "selected_success_at_8": 0.0,
1422
+ "selected_success_gain_over_base_at_8": 0.0,
1423
+ "selected_unsafe_known_at_8": 0.0,
1424
+ "selected_utility_at_8": 0.1222188975661993,
1425
+ "selected_utility_gain_over_base_at_8": -0.024850530549883842,
1426
+ "selector_gap_fraction_to_hidden_at_8": 3.5662633342346166,
1427
+ "selector_regret_at_8": 0.02600152976810932,
1428
+ "success_selector_gap_at_8": 0.0,
1429
+ "success_support_gap_at_8": 0.0,
1430
+ "success_total_car_to_hidden_at_8": 0.0,
1431
+ "support_gap_at_8": 0.08758716285228729,
1432
+ "support_gap_fraction_to_hidden_at_8": 1.41259115683824,
1433
+ "total_car_to_hidden_at_8": 0.11358869262039661
1434
+ },
1435
+ "75070641": {
1436
+ "any_generated_unsafe_known_at_8": 0.0,
1437
+ "base_safety_label_known": 1.0,
1438
+ "base_success": 0.0,
1439
+ "base_unsafe_known": 0.0,
1440
+ "base_utility": 0.194026380777359,
1441
+ "branch_car_at_8": 0.024178534746170044,
1442
+ "generated_safety_label_coverage_at_8": 1.0,
1443
+ "generated_unsafe_rate_known_at_8": 0.0,
1444
+ "hidden_chart_oracle_success_at_8": 1.0,
1445
+ "hidden_chart_oracle_utility_at_8": 2.0,
1446
+ "ncar_to_hidden_chart_oracle_at_8": 1.0008596290963836,
1447
+ "ncar_to_proposal_oracle_at_8": 1.0686141104703737,
1448
+ "outcome_ptr_at_8": 1.0,
1449
+ "pairwise_causal_calibration_ece": 0.1631929231601434,
1450
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1451
+ "proposal_oracle_success_at_8": 0.0,
1452
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1453
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1454
+ "proposal_oracle_utility_at_8": 0.21665243804454803,
1455
+ "proposal_oracle_utility_gain_over_base_at_8": 0.022626057267189026,
1456
+ "selected_safety_label_known_at_8": 1.0,
1457
+ "selected_success_at_8": 0.0,
1458
+ "selected_success_gain_over_base_at_8": 0.0,
1459
+ "selected_unsafe_known_at_8": 0.0,
1460
+ "selected_utility_at_8": 0.192473903298378,
1461
+ "selected_utility_gain_over_base_at_8": -0.001552477478981018,
1462
+ "selector_gap_fraction_to_hidden_at_8": 0.01338808855722787,
1463
+ "selector_regret_at_8": 0.024178534746170044,
1464
+ "success_selector_gap_at_8": 0.0,
1465
+ "success_support_gap_at_8": 1.0,
1466
+ "success_total_car_to_hidden_at_8": 1.0,
1467
+ "support_gap_at_8": 1.783347561955452,
1468
+ "support_gap_fraction_to_hidden_at_8": 0.987471546081095,
1469
+ "total_car_to_hidden_at_8": 1.807526096701622
1470
+ },
1471
+ "767747424": {
1472
+ "any_generated_unsafe_known_at_8": 0.5,
1473
+ "base_safety_label_known": 1.0,
1474
+ "base_success": 0.0,
1475
+ "base_unsafe_known": 0.0,
1476
+ "base_utility": 0.10894576646387577,
1477
+ "branch_car_at_8": 0.016496962402015924,
1478
+ "generated_safety_label_coverage_at_8": 1.0,
1479
+ "generated_unsafe_rate_known_at_8": 0.5,
1480
+ "hidden_chart_oracle_success_at_8": 0.0,
1481
+ "hidden_chart_oracle_utility_at_8": 0.18863565661013126,
1482
+ "ncar_to_hidden_chart_oracle_at_8": 5.205313661372345,
1483
+ "ncar_to_proposal_oracle_at_8": 0.9043237597220475,
1484
+ "outcome_ptr_at_8": 0.5,
1485
+ "pairwise_causal_calibration_ece": 0.24962611487178488,
1486
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1487
+ "proposal_oracle_success_at_8": 0.0,
1488
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1489
+ "proposal_oracle_unsafe_known_at_8": 0.5,
1490
+ "proposal_oracle_utility_at_8": 0.10893216636031866,
1491
+ "proposal_oracle_utility_gain_over_base_at_8": -1.3600103557109833e-05,
1492
+ "selected_safety_label_known_at_8": 1.0,
1493
+ "selected_success_at_8": 0.0,
1494
+ "selected_success_gain_over_base_at_8": 0.0,
1495
+ "selected_unsafe_known_at_8": 0.5,
1496
+ "selected_utility_at_8": 0.09243520395830274,
1497
+ "selected_utility_gain_over_base_at_8": -0.016510562505573034,
1498
+ "selector_gap_fraction_to_hidden_at_8": 2.999421934457336,
1499
+ "selector_regret_at_8": 0.016496962402015924,
1500
+ "success_selector_gap_at_8": 0.0,
1501
+ "success_support_gap_at_8": 0.0,
1502
+ "success_total_car_to_hidden_at_8": 0.0,
1503
+ "support_gap_at_8": 0.0797034902498126,
1504
+ "support_gap_fraction_to_hidden_at_8": 2.2059097841347444,
1505
+ "total_car_to_hidden_at_8": 0.09620045265182853
1506
+ },
1507
+ "77147264": {
1508
+ "any_generated_unsafe_known_at_8": 0.0,
1509
+ "base_safety_label_known": 1.0,
1510
+ "base_success": 0.5,
1511
+ "base_unsafe_known": 0.0,
1512
+ "base_utility": 1.0903030335903168,
1513
+ "branch_car_at_8": 0.00890406221151352,
1514
+ "generated_safety_label_coverage_at_8": 1.0,
1515
+ "generated_unsafe_rate_known_at_8": 0.0,
1516
+ "hidden_chart_oracle_success_at_8": 1.0,
1517
+ "hidden_chart_oracle_utility_at_8": 2.0,
1518
+ "ncar_to_hidden_chart_oracle_at_8": 1.0010257406519596,
1519
+ "ncar_to_proposal_oracle_at_8": 1.1170642490674998,
1520
+ "outcome_ptr_at_8": 0.5,
1521
+ "pairwise_causal_calibration_ece": 0.05996298782788377,
1522
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1523
+ "proposal_oracle_success_at_8": 0.5,
1524
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1525
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1526
+ "proposal_oracle_utility_at_8": 1.098273977637291,
1527
+ "proposal_oracle_utility_gain_over_base_at_8": 0.007970944046974182,
1528
+ "selected_safety_label_known_at_8": 1.0,
1529
+ "selected_success_at_8": 0.5,
1530
+ "selected_success_gain_over_base_at_8": 0.0,
1531
+ "selected_unsafe_known_at_8": 0.0,
1532
+ "selected_utility_at_8": 1.0893699154257774,
1533
+ "selected_utility_gain_over_base_at_8": -0.0009331181645393372,
1534
+ "selector_gap_fraction_to_hidden_at_8": 0.009787943172609816,
1535
+ "selector_regret_at_8": 0.00890406221151352,
1536
+ "success_selector_gap_at_8": 0.0,
1537
+ "success_support_gap_at_8": 0.5,
1538
+ "success_total_car_to_hidden_at_8": 0.5,
1539
+ "support_gap_at_8": 0.901726022362709,
1540
+ "support_gap_fraction_to_hidden_at_8": 0.9912378029813232,
1541
+ "total_car_to_hidden_at_8": 0.9106300845742226
1542
+ },
1543
+ "794181626": {
1544
+ "any_generated_unsafe_known_at_8": 0.0,
1545
+ "base_safety_label_known": 1.0,
1546
+ "base_success": 1.0,
1547
+ "base_unsafe_known": 0.0,
1548
+ "base_utility": 2.0,
1549
+ "branch_car_at_8": 0.0,
1550
+ "generated_safety_label_coverage_at_8": 1.0,
1551
+ "generated_unsafe_rate_known_at_8": 0.0,
1552
+ "hidden_chart_oracle_success_at_8": 1.0,
1553
+ "hidden_chart_oracle_utility_at_8": 2.0,
1554
+ "outcome_ptr_at_8": 0.0,
1555
+ "pairwise_causal_calibration_ece": 0.0,
1556
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1557
+ "proposal_oracle_success_at_8": 1.0,
1558
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1559
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1560
+ "proposal_oracle_utility_at_8": 2.0,
1561
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
1562
+ "selected_safety_label_known_at_8": 1.0,
1563
+ "selected_success_at_8": 1.0,
1564
+ "selected_success_gain_over_base_at_8": 0.0,
1565
+ "selected_unsafe_known_at_8": 0.0,
1566
+ "selected_utility_at_8": 2.0,
1567
+ "selected_utility_gain_over_base_at_8": 0.0,
1568
+ "selector_regret_at_8": 0.0,
1569
+ "success_selector_gap_at_8": 0.0,
1570
+ "success_support_gap_at_8": 0.0,
1571
+ "success_total_car_to_hidden_at_8": 0.0,
1572
+ "support_gap_at_8": 0.0,
1573
+ "total_car_to_hidden_at_8": 0.0
1574
+ },
1575
+ "957101489": {
1576
+ "any_generated_unsafe_known_at_8": 0.0,
1577
+ "base_safety_label_known": 1.0,
1578
+ "base_success": 0.0,
1579
+ "base_unsafe_known": 0.0,
1580
+ "base_utility": 0.23217551410198212,
1581
+ "branch_car_at_8": 0.3351810723543167,
1582
+ "generated_safety_label_coverage_at_8": 1.0,
1583
+ "generated_unsafe_rate_known_at_8": 0.0,
1584
+ "hidden_chart_oracle_success_at_8": 1.0,
1585
+ "hidden_chart_oracle_utility_at_8": 2.0,
1586
+ "ncar_to_hidden_chart_oracle_at_8": 0.998510000020915,
1587
+ "ncar_to_proposal_oracle_at_8": 0.9922026628719985,
1588
+ "outcome_ptr_at_8": 1.0,
1589
+ "pairwise_causal_calibration_ece": 0.4692653473226291,
1590
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1591
+ "proposal_oracle_success_at_8": 0.0,
1592
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1593
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1594
+ "proposal_oracle_utility_at_8": 0.5699906349182129,
1595
+ "proposal_oracle_utility_gain_over_base_at_8": 0.3378151208162308,
1596
+ "selected_safety_label_known_at_8": 1.0,
1597
+ "selected_success_at_8": 0.0,
1598
+ "selected_success_gain_over_base_at_8": 0.0,
1599
+ "selected_unsafe_known_at_8": 0.0,
1600
+ "selected_utility_at_8": 0.23480956256389618,
1601
+ "selected_utility_gain_over_base_at_8": 0.0026340484619140625,
1602
+ "selector_gap_fraction_to_hidden_at_8": 0.18960087668660824,
1603
+ "selector_regret_at_8": 0.3351810723543167,
1604
+ "success_selector_gap_at_8": 0.0,
1605
+ "success_support_gap_at_8": 1.0,
1606
+ "success_total_car_to_hidden_at_8": 1.0,
1607
+ "support_gap_at_8": 1.430009365081787,
1608
+ "support_gap_fraction_to_hidden_at_8": 0.8089091289825484,
1609
+ "total_car_to_hidden_at_8": 1.7651904374361038
1610
+ },
1611
+ "998761162": {
1612
+ "any_generated_unsafe_known_at_8": 0.0,
1613
+ "base_safety_label_known": 1.0,
1614
+ "base_success": 0.0,
1615
+ "base_unsafe_known": 0.0,
1616
+ "base_utility": 0.20425210893154144,
1617
+ "branch_car_at_8": 0.004930347204208374,
1618
+ "generated_safety_label_coverage_at_8": 1.0,
1619
+ "generated_unsafe_rate_known_at_8": 0.0,
1620
+ "hidden_chart_oracle_success_at_8": 1.0,
1621
+ "hidden_chart_oracle_utility_at_8": 2.0,
1622
+ "ncar_to_hidden_chart_oracle_at_8": 1.062306753304616,
1623
+ "ncar_to_proposal_oracle_at_8": 0.04609658160039006,
1624
+ "outcome_ptr_at_8": 0.0,
1625
+ "pairwise_causal_calibration_ece": 0.24317634255647042,
1626
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1627
+ "proposal_oracle_success_at_8": 0.0,
1628
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1629
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1630
+ "proposal_oracle_utility_at_8": 0.09729522466659546,
1631
+ "proposal_oracle_utility_gain_over_base_at_8": -0.10695688426494598,
1632
+ "selected_safety_label_known_at_8": 1.0,
1633
+ "selected_success_at_8": 0.0,
1634
+ "selected_success_gain_over_base_at_8": 0.0,
1635
+ "selected_unsafe_known_at_8": 0.0,
1636
+ "selected_utility_at_8": 0.09236487746238708,
1637
+ "selected_utility_gain_over_base_at_8": -0.11188723146915436,
1638
+ "selector_gap_fraction_to_hidden_at_8": 0.002745567587037426,
1639
+ "selector_regret_at_8": 0.004930347204208374,
1640
+ "success_selector_gap_at_8": 0.0,
1641
+ "success_support_gap_at_8": 1.0,
1642
+ "success_total_car_to_hidden_at_8": 1.0,
1643
+ "support_gap_at_8": 1.9027047753334045,
1644
+ "support_gap_fraction_to_hidden_at_8": 1.0595611916332572,
1645
+ "total_car_to_hidden_at_8": 1.907635122537613
1646
+ }
1647
+ }
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/measured_metrics/metrics_by_task.json ADDED
@@ -0,0 +1,182 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "LiftPegUpright-v1": {
3
+ "any_generated_unsafe_known_at_8": 0.0,
4
+ "base_safety_label_known": 1.0,
5
+ "base_success": 0.0,
6
+ "base_unsafe_known": 0.0,
7
+ "base_utility": 0.40346155563990277,
8
+ "branch_car_at_8": 0.217838483552138,
9
+ "generated_safety_label_coverage_at_8": 1.0,
10
+ "generated_unsafe_rate_known_at_8": 0.0,
11
+ "hidden_chart_oracle_success_at_8": 1.0,
12
+ "hidden_chart_oracle_utility_at_8": 2.0,
13
+ "ncar_to_hidden_chart_oracle_at_8": 1.0077789117032718,
14
+ "ncar_to_proposal_oracle_at_8": 0.5175964402383324,
15
+ "outcome_ptr_at_8": 0.5,
16
+ "pairwise_causal_calibration_ece": 0.2108209521734508,
17
+ "proposal_oracle_safety_label_known_at_8": 1.0,
18
+ "proposal_oracle_success_at_8": 0.16666666666666666,
19
+ "proposal_oracle_success_gain_over_base_at_8": 0.16666666666666666,
20
+ "proposal_oracle_unsafe_known_at_8": 0.0,
21
+ "proposal_oracle_utility_at_8": 0.6081412807106972,
22
+ "proposal_oracle_utility_gain_over_base_at_8": 0.20467972507079443,
23
+ "selected_safety_label_known_at_8": 1.0,
24
+ "selected_success_at_8": 0.0,
25
+ "selected_success_gain_over_base_at_8": 0.0,
26
+ "selected_unsafe_known_at_8": 0.0,
27
+ "selected_utility_at_8": 0.39030279715855914,
28
+ "selected_utility_gain_over_base_at_8": -0.013158758481343588,
29
+ "selector_gap_fraction_to_hidden_at_8": 0.1701968825815193,
30
+ "selector_regret_at_8": 0.217838483552138,
31
+ "success_selector_gap_at_8": 0.16666666666666666,
32
+ "success_support_gap_at_8": 0.8333333333333334,
33
+ "success_total_car_to_hidden_at_8": 1.0,
34
+ "support_gap_at_8": 1.3918587192893028,
35
+ "support_gap_fraction_to_hidden_at_8": 0.8375820355789246,
36
+ "total_car_to_hidden_at_8": 1.609697202841441
37
+ },
38
+ "PickCube-v1": {
39
+ "any_generated_unsafe_known_at_8": 0.0,
40
+ "base_safety_label_known": 1.0,
41
+ "base_success": 0.0,
42
+ "base_unsafe_known": 0.0,
43
+ "base_utility": 0.11945180594921112,
44
+ "branch_car_at_8": 0.011674322187900543,
45
+ "generated_safety_label_coverage_at_8": 1.0,
46
+ "generated_unsafe_rate_known_at_8": 0.0,
47
+ "hidden_chart_oracle_success_at_8": 1.0,
48
+ "hidden_chart_oracle_utility_at_8": 2.0,
49
+ "ncar_to_hidden_chart_oracle_at_8": 1.0157636165911068,
50
+ "ncar_to_proposal_oracle_at_8": 0.649658432039702,
51
+ "outcome_ptr_at_8": 0.0,
52
+ "pairwise_causal_calibration_ece": 0.11483608299984571,
53
+ "proposal_oracle_safety_label_known_at_8": 1.0,
54
+ "proposal_oracle_success_at_8": 0.0,
55
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
56
+ "proposal_oracle_unsafe_known_at_8": 0.0,
57
+ "proposal_oracle_utility_at_8": 0.10148187726736069,
58
+ "proposal_oracle_utility_gain_over_base_at_8": -0.017969928681850433,
59
+ "selected_safety_label_known_at_8": 1.0,
60
+ "selected_success_at_8": 0.0,
61
+ "selected_success_gain_over_base_at_8": 0.0,
62
+ "selected_unsafe_known_at_8": 0.0,
63
+ "selected_utility_at_8": 0.08980755507946014,
64
+ "selected_utility_gain_over_base_at_8": -0.029644250869750977,
65
+ "selector_gap_fraction_to_hidden_at_8": 0.006207935656652067,
66
+ "selector_regret_at_8": 0.011674322187900543,
67
+ "success_selector_gap_at_8": 0.0,
68
+ "success_support_gap_at_8": 1.0,
69
+ "success_total_car_to_hidden_at_8": 1.0,
70
+ "support_gap_at_8": 1.8985181227326393,
71
+ "support_gap_fraction_to_hidden_at_8": 1.0095556863358777,
72
+ "total_car_to_hidden_at_8": 1.9101924449205399
73
+ },
74
+ "PullCube-v1": {
75
+ "any_generated_unsafe_known_at_8": 0.6764705882352942,
76
+ "base_safety_label_known": 1.0,
77
+ "base_success": 0.20588235294117646,
78
+ "base_unsafe_known": 0.0,
79
+ "base_utility": 0.5097779550534838,
80
+ "branch_car_at_8": 0.06663480061380302,
81
+ "generated_safety_label_coverage_at_8": 1.0,
82
+ "generated_unsafe_rate_known_at_8": 0.5330882352941176,
83
+ "hidden_chart_oracle_success_at_8": 0.3235294117647059,
84
+ "hidden_chart_oracle_utility_at_8": 0.7823218367993832,
85
+ "ncar_to_hidden_chart_oracle_at_8": 2.7805691722759143,
86
+ "ncar_to_proposal_oracle_at_8": 0.6898506880563886,
87
+ "outcome_ptr_at_8": 0.2647058823529412,
88
+ "pairwise_causal_calibration_ece": 0.21534074294224181,
89
+ "proposal_oracle_safety_label_known_at_8": 1.0,
90
+ "proposal_oracle_success_at_8": 0.08823529411764706,
91
+ "proposal_oracle_success_gain_over_base_at_8": -0.11764705882352941,
92
+ "proposal_oracle_unsafe_known_at_8": 0.5,
93
+ "proposal_oracle_utility_at_8": 0.27341561364557815,
94
+ "proposal_oracle_utility_gain_over_base_at_8": -0.2363623414079056,
95
+ "selected_safety_label_known_at_8": 1.0,
96
+ "selected_success_at_8": 0.058823529411764705,
97
+ "selected_success_gain_over_base_at_8": -0.14705882352941177,
98
+ "selected_unsafe_known_at_8": 0.4411764705882353,
99
+ "selected_utility_at_8": 0.20678081303177512,
100
+ "selected_utility_gain_over_base_at_8": -0.30299714202170863,
101
+ "selector_gap_fraction_to_hidden_at_8": 0.9297287801367955,
102
+ "selector_regret_at_8": 0.06663480061380302,
103
+ "success_selector_gap_at_8": 0.029411764705882353,
104
+ "success_support_gap_at_8": 0.23529411764705882,
105
+ "success_total_car_to_hidden_at_8": 0.2647058823529412,
106
+ "support_gap_at_8": 0.5125939046744915,
107
+ "support_gap_fraction_to_hidden_at_8": 1.9391989570475572,
108
+ "total_car_to_hidden_at_8": 0.5767036846784108
109
+ },
110
+ "PushCube-v1": {
111
+ "any_generated_unsafe_known_at_8": 0.0,
112
+ "base_safety_label_known": 1.0,
113
+ "base_success": 0.6923076923076923,
114
+ "base_unsafe_known": 0.0,
115
+ "base_utility": 1.4275685124672377,
116
+ "branch_car_at_8": 0.1499279118501223,
117
+ "generated_safety_label_coverage_at_8": 1.0,
118
+ "generated_unsafe_rate_known_at_8": 0.0,
119
+ "hidden_chart_oracle_success_at_8": 1.0,
120
+ "hidden_chart_oracle_utility_at_8": 2.0,
121
+ "ncar_to_hidden_chart_oracle_at_8": 1.0314368776438072,
122
+ "ncar_to_proposal_oracle_at_8": 0.26408516243539404,
123
+ "outcome_ptr_at_8": 0.0,
124
+ "pairwise_causal_calibration_ece": 0.1875472363576914,
125
+ "proposal_oracle_safety_label_known_at_8": 1.0,
126
+ "proposal_oracle_success_at_8": 0.46153846153846156,
127
+ "proposal_oracle_success_gain_over_base_at_8": -0.23076923076923078,
128
+ "proposal_oracle_unsafe_known_at_8": 0.0,
129
+ "proposal_oracle_utility_at_8": 0.9768703006781064,
130
+ "proposal_oracle_utility_gain_over_base_at_8": -0.45069821178913116,
131
+ "selected_safety_label_known_at_8": 1.0,
132
+ "selected_success_at_8": 0.38461538461538464,
133
+ "selected_success_gain_over_base_at_8": -0.3076923076923077,
134
+ "selected_unsafe_known_at_8": 0.0,
135
+ "selected_utility_at_8": 0.8269423888279841,
136
+ "selected_utility_gain_over_base_at_8": -0.6006261236392535,
137
+ "selector_gap_fraction_to_hidden_at_8": 0.007570377855345064,
138
+ "selector_regret_at_8": 0.1499279118501223,
139
+ "success_selector_gap_at_8": 0.07692307692307693,
140
+ "success_support_gap_at_8": 0.5384615384615384,
141
+ "success_total_car_to_hidden_at_8": 0.6153846153846154,
142
+ "support_gap_at_8": 1.0231296993218935,
143
+ "support_gap_fraction_to_hidden_at_8": 1.0238665053485367,
144
+ "total_car_to_hidden_at_8": 1.173057611172016
145
+ },
146
+ "StackCube-v1": {
147
+ "any_generated_unsafe_known_at_8": 0.0,
148
+ "base_safety_label_known": 1.0,
149
+ "base_success": 0.2,
150
+ "base_unsafe_known": 0.0,
151
+ "base_utility": 0.696571855743726,
152
+ "branch_car_at_8": 0.10400179326534272,
153
+ "generated_safety_label_coverage_at_8": 1.0,
154
+ "generated_unsafe_rate_known_at_8": 0.0,
155
+ "hidden_chart_oracle_success_at_8": 1.0,
156
+ "hidden_chart_oracle_utility_at_8": 2.0,
157
+ "ncar_to_hidden_chart_oracle_at_8": 1.002923757801857,
158
+ "ncar_to_proposal_oracle_at_8": 3.1742031183267274,
159
+ "outcome_ptr_at_8": 0.7333333333333333,
160
+ "pairwise_causal_calibration_ece": 0.20429837578002458,
161
+ "proposal_oracle_safety_label_known_at_8": 1.0,
162
+ "proposal_oracle_success_at_8": 0.06666666666666667,
163
+ "proposal_oracle_success_gain_over_base_at_8": -0.13333333333333333,
164
+ "proposal_oracle_unsafe_known_at_8": 0.0,
165
+ "proposal_oracle_utility_at_8": 0.5997182667255402,
166
+ "proposal_oracle_utility_gain_over_base_at_8": -0.09685358901818593,
167
+ "selected_safety_label_known_at_8": 1.0,
168
+ "selected_success_at_8": 0.06666666666666667,
169
+ "selected_success_gain_over_base_at_8": -0.13333333333333333,
170
+ "selected_unsafe_known_at_8": 0.0,
171
+ "selected_utility_at_8": 0.4957164734601974,
172
+ "selected_utility_gain_over_base_at_8": -0.20085538228352864,
173
+ "selector_gap_fraction_to_hidden_at_8": 0.04533697018458092,
174
+ "selector_regret_at_8": 0.10400179326534272,
175
+ "success_selector_gap_at_8": 0.0,
176
+ "success_support_gap_at_8": 0.9333333333333333,
177
+ "success_total_car_to_hidden_at_8": 0.9333333333333333,
178
+ "support_gap_at_8": 1.40028173327446,
179
+ "support_gap_fraction_to_hidden_at_8": 0.9575867940016279,
180
+ "total_car_to_hidden_at_8": 1.5042835265398025
181
+ }
182
+ }
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/measured_metrics/report.md ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Metric Evaluation (measured)
2
+
3
+ K: `8`
4
+
5
+ | Metric | N | Micro mean | 95% CI | Task macro | Seed macro |
6
+ | --- | ---: | ---: | ---: | ---: | ---: |
7
+ | any_generated_unsafe_known_at_8 | 69 | 0.3333 | [0.2319, 0.4493] | 0.1353 | 0.2227 |
8
+ | base_safety_label_known | 69 | 1.0000 | [1.0000, 1.0000] | 1.0000 | 1.0000 |
9
+ | base_success | 69 | 0.2754 | [0.1739, 0.3768] | 0.2196 | 0.3216 |
10
+ | base_unsafe_known | 69 | 0.0000 | [0.0000, 0.0000] | 0.0000 | 0.0000 |
11
+ | base_utility | 69 | 0.7084 | [0.5356, 0.8955] | 0.6314 | 0.8273 |
12
+ | branch_car_at_8 | 69 | 0.1028 | [0.0393, 0.1788] | 0.1100 | 0.1188 |
13
+ | generated_safety_label_coverage_at_8 | 69 | 1.0000 | [1.0000, 1.0000] | 1.0000 | 1.0000 |
14
+ | generated_unsafe_rate_known_at_8 | 69 | 0.2627 | [0.1757, 0.3587] | 0.1066 | 0.1807 |
15
+ | hidden_chart_oracle_success_at_8 | 69 | 0.6667 | [0.5507, 0.7826] | 0.8647 | 0.8358 |
16
+ | hidden_chart_oracle_utility_at_8 | 69 | 1.4000 | [1.1926, 1.6081] | 1.7565 | 1.7062 |
17
+ | ncar_to_hidden_chart_oracle_at_8 | 50 | 1.9660 | [1.3091, 2.9683] | 1.3677 | 1.5829 |
18
+ | ncar_to_proposal_oracle_at_8 | 58 | 1.2226 | [0.4440, 2.6378] | 1.0591 | 1.4953 |
19
+ | outcome_ptr_at_8 | 69 | 0.3333 | [0.2174, 0.4348] | 0.2996 | 0.3330 |
20
+ | pairwise_causal_calibration_ece | 69 | 0.2059 | [0.1648, 0.2356] | 0.1866 | 0.2013 |
21
+ | proposal_oracle_safety_label_known_at_8 | 69 | 1.0000 | [1.0000, 1.0000] | 1.0000 | 1.0000 |
22
+ | proposal_oracle_success_at_8 | 69 | 0.1594 | [0.0870, 0.2609] | 0.1566 | 0.1986 |
23
+ | proposal_oracle_success_gain_over_base_at_8 | 69 | -0.1159 | [-0.1884, -0.0435] | -0.0630 | -0.1230 |
24
+ | proposal_oracle_unsafe_known_at_8 | 69 | 0.2464 | [0.1739, 0.3478] | 0.1000 | 0.1670 |
25
+ | proposal_oracle_utility_at_8 | 69 | 0.5035 | [0.3582, 0.6941] | 0.5119 | 0.6176 |
26
+ | proposal_oracle_utility_gain_over_base_at_8 | 69 | -0.2049 | [-0.3325, -0.0815] | -0.1194 | -0.2098 |
27
+ | selected_safety_label_known_at_8 | 69 | 1.0000 | [1.0000, 1.0000] | 1.0000 | 1.0000 |
28
+ | selected_success_at_8 | 69 | 0.1159 | [0.0580, 0.2174] | 0.1020 | 0.1489 |
29
+ | selected_success_gain_over_base_at_8 | 69 | -0.1594 | [-0.2464, -0.0870] | -0.1176 | -0.1727 |
30
+ | selected_unsafe_known_at_8 | 69 | 0.2174 | [0.1449, 0.3043] | 0.0882 | 0.1557 |
31
+ | selected_utility_at_8 | 69 | 0.4007 | [0.2624, 0.5763] | 0.4019 | 0.4988 |
32
+ | selected_utility_gain_over_base_at_8 | 69 | -0.3077 | [-0.4717, -0.1684] | -0.2295 | -0.3285 |
33
+ | selector_gap_fraction_to_hidden_at_8 | 50 | 0.5341 | [0.1311, 1.1724] | 0.2318 | 0.3003 |
34
+ | selector_regret_at_8 | 69 | 0.1028 | [0.0393, 0.1788] | 0.1100 | 0.1188 |
35
+ | success_selector_gap_at_8 | 69 | 0.0435 | [0.0145, 0.0870] | 0.0546 | 0.0496 |
36
+ | success_support_gap_at_8 | 69 | 0.5072 | [0.3768, 0.6232] | 0.7081 | 0.6372 |
37
+ | success_total_car_to_hidden_at_8 | 69 | 0.5507 | [0.4203, 0.6812] | 0.7627 | 0.6869 |
38
+ | support_gap_at_8 | 69 | 0.8983 | [0.6788, 1.0761] | 1.2453 | 1.0894 |
39
+ | support_gap_fraction_to_hidden_at_8 | 50 | 1.4796 | [1.1215, 1.9117] | 1.1536 | 1.3055 |
40
+ | total_car_to_hidden_at_8 | 69 | 0.9999 | [0.7623, 1.2048] | 1.3548 | 1.2076 |
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/measured_metrics/table.tex ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ % Auto-generated by scripts/eval_metrics.py
2
+ \begin{tabular}{lrrrr}
3
+ \toprule
4
+ Metric & N & Micro mean & CI low & CI high \\
5
+ \midrule
6
+ any\_generated\_unsafe\_known\_at\_8 & 69 & 0.3333 & 0.2319 & 0.4493 \\
7
+ base\_safety\_label\_known & 69 & 1.0000 & 1.0000 & 1.0000 \\
8
+ base\_success & 69 & 0.2754 & 0.1739 & 0.3768 \\
9
+ base\_unsafe\_known & 69 & 0.0000 & 0.0000 & 0.0000 \\
10
+ base\_utility & 69 & 0.7084 & 0.5356 & 0.8955 \\
11
+ branch\_car\_at\_8 & 69 & 0.1028 & 0.0393 & 0.1788 \\
12
+ generated\_safety\_label\_coverage\_at\_8 & 69 & 1.0000 & 1.0000 & 1.0000 \\
13
+ generated\_unsafe\_rate\_known\_at\_8 & 69 & 0.2627 & 0.1757 & 0.3587 \\
14
+ hidden\_chart\_oracle\_success\_at\_8 & 69 & 0.6667 & 0.5507 & 0.7826 \\
15
+ hidden\_chart\_oracle\_utility\_at\_8 & 69 & 1.4000 & 1.1926 & 1.6081 \\
16
+ ncar\_to\_hidden\_chart\_oracle\_at\_8 & 50 & 1.9660 & 1.3091 & 2.9683 \\
17
+ ncar\_to\_proposal\_oracle\_at\_8 & 58 & 1.2226 & 0.4440 & 2.6378 \\
18
+ outcome\_ptr\_at\_8 & 69 & 0.3333 & 0.2174 & 0.4348 \\
19
+ pairwise\_causal\_calibration\_ece & 69 & 0.2059 & 0.1648 & 0.2356 \\
20
+ proposal\_oracle\_safety\_label\_known\_at\_8 & 69 & 1.0000 & 1.0000 & 1.0000 \\
21
+ proposal\_oracle\_success\_at\_8 & 69 & 0.1594 & 0.0870 & 0.2609 \\
22
+ proposal\_oracle\_success\_gain\_over\_base\_at\_8 & 69 & -0.1159 & -0.1884 & -0.0435 \\
23
+ proposal\_oracle\_unsafe\_known\_at\_8 & 69 & 0.2464 & 0.1739 & 0.3478 \\
24
+ proposal\_oracle\_utility\_at\_8 & 69 & 0.5035 & 0.3582 & 0.6941 \\
25
+ proposal\_oracle\_utility\_gain\_over\_base\_at\_8 & 69 & -0.2049 & -0.3325 & -0.0815 \\
26
+ selected\_safety\_label\_known\_at\_8 & 69 & 1.0000 & 1.0000 & 1.0000 \\
27
+ selected\_success\_at\_8 & 69 & 0.1159 & 0.0580 & 0.2174 \\
28
+ selected\_success\_gain\_over\_base\_at\_8 & 69 & -0.1594 & -0.2464 & -0.0870 \\
29
+ selected\_unsafe\_known\_at\_8 & 69 & 0.2174 & 0.1449 & 0.3043 \\
30
+ selected\_utility\_at\_8 & 69 & 0.4007 & 0.2624 & 0.5763 \\
31
+ selected\_utility\_gain\_over\_base\_at\_8 & 69 & -0.3077 & -0.4717 & -0.1684 \\
32
+ selector\_gap\_fraction\_to\_hidden\_at\_8 & 50 & 0.5341 & 0.1311 & 1.1724 \\
33
+ selector\_regret\_at\_8 & 69 & 0.1028 & 0.0393 & 0.1788 \\
34
+ success\_selector\_gap\_at\_8 & 69 & 0.0435 & 0.0145 & 0.0870 \\
35
+ success\_support\_gap\_at\_8 & 69 & 0.5072 & 0.3768 & 0.6232 \\
36
+ success\_total\_car\_to\_hidden\_at\_8 & 69 & 0.5507 & 0.4203 & 0.6812 \\
37
+ support\_gap\_at\_8 & 69 & 0.8983 & 0.6788 & 1.0761 \\
38
+ support\_gap\_fraction\_to\_hidden\_at\_8 & 50 & 1.4796 & 1.1215 & 1.9117 \\
39
+ total\_car\_to\_hidden\_at\_8 & 69 & 0.9999 & 0.7623 & 1.2048 \\
40
+ \bottomrule
41
+ \end{tabular}
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/metrics.json ADDED
@@ -0,0 +1,3041 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "checkpoint": "runs/ctt_residual_base_context_obs_seed2/model.pt",
3
+ "k": 8,
4
+ "measured_metric_summary": {
5
+ "any_generated_unsafe_known_at_8": {
6
+ "macro_by_seed": {
7
+ "ci": {
8
+ "confidence": 0.95,
9
+ "high": 0.34255319148936175,
10
+ "low": 0.11985815602836879,
11
+ "mean": 0.22269503546099287,
12
+ "n": 47,
13
+ "num_samples": 200
14
+ },
15
+ "mean": 0.22269503546099287,
16
+ "per_seed": {
17
+ "1025740212": 0.0,
18
+ "1055157354": 0.0,
19
+ "1055567618": 0.0,
20
+ "1113268940": 0.0,
21
+ "1113371261": 0.0,
22
+ "1176943004": 0.0,
23
+ "1189392917": 0.0,
24
+ "1231780997": 0.0,
25
+ "1271718007": 0.0,
26
+ "1273247072": 0.0,
27
+ "1358421325": 0.0,
28
+ "1375492786": 1.0,
29
+ "141984615": 0.3333333333333333,
30
+ "1475440973": 0.0,
31
+ "1483645336": 0.0,
32
+ "1522272686": 0.0,
33
+ "1528740417": 0.0,
34
+ "1549324805": 1.0,
35
+ "1607773860": 0.0,
36
+ "1665739511": 0.0,
37
+ "1747178030": 1.0,
38
+ "1881223807": 0.0,
39
+ "1919198188": 0.0,
40
+ "1991414747": 0.0,
41
+ "1997090716": 0.0,
42
+ "2015956957": 0.8,
43
+ "2054899126": 1.0,
44
+ "211245842": 0.0,
45
+ "2141802816": 0.0,
46
+ "256770874": 0.0,
47
+ "278008524": 0.5,
48
+ "416183106": 0.0,
49
+ "534717141": 0.0,
50
+ "551504635": 0.0,
51
+ "565369392": 1.0,
52
+ "565621826": 1.0,
53
+ "585904130": 0.3333333333333333,
54
+ "626486813": 0.5,
55
+ "636680137": 1.0,
56
+ "706479600": 0.0,
57
+ "750448296": 0.5,
58
+ "75070641": 0.0,
59
+ "767747424": 0.5,
60
+ "77147264": 0.0,
61
+ "794181626": 0.0,
62
+ "957101489": 0.0,
63
+ "998761162": 0.0
64
+ }
65
+ },
66
+ "macro_by_task": {
67
+ "ci": {
68
+ "confidence": 0.95,
69
+ "high": 0.40588235294117647,
70
+ "low": 0.0,
71
+ "mean": 0.13529411764705884,
72
+ "n": 5,
73
+ "num_samples": 200
74
+ },
75
+ "mean": 0.13529411764705884,
76
+ "per_task": {
77
+ "LiftPegUpright-v1": 0.0,
78
+ "PickCube-v1": 0.0,
79
+ "PullCube-v1": 0.6764705882352942,
80
+ "PushCube-v1": 0.0,
81
+ "StackCube-v1": 0.0
82
+ }
83
+ },
84
+ "metric": "any_generated_unsafe_known_at_8",
85
+ "micro": {
86
+ "confidence": 0.95,
87
+ "high": 0.4492753623188406,
88
+ "low": 0.2318840579710145,
89
+ "mean": 0.3333333333333333,
90
+ "n": 69,
91
+ "num_samples": 200
92
+ },
93
+ "n": 69
94
+ },
95
+ "base_safety_label_known": {
96
+ "macro_by_seed": {
97
+ "ci": {
98
+ "confidence": 0.95,
99
+ "high": 1.0,
100
+ "low": 1.0,
101
+ "mean": 1.0,
102
+ "n": 47,
103
+ "num_samples": 200
104
+ },
105
+ "mean": 1.0,
106
+ "per_seed": {
107
+ "1025740212": 1.0,
108
+ "1055157354": 1.0,
109
+ "1055567618": 1.0,
110
+ "1113268940": 1.0,
111
+ "1113371261": 1.0,
112
+ "1176943004": 1.0,
113
+ "1189392917": 1.0,
114
+ "1231780997": 1.0,
115
+ "1271718007": 1.0,
116
+ "1273247072": 1.0,
117
+ "1358421325": 1.0,
118
+ "1375492786": 1.0,
119
+ "141984615": 1.0,
120
+ "1475440973": 1.0,
121
+ "1483645336": 1.0,
122
+ "1522272686": 1.0,
123
+ "1528740417": 1.0,
124
+ "1549324805": 1.0,
125
+ "1607773860": 1.0,
126
+ "1665739511": 1.0,
127
+ "1747178030": 1.0,
128
+ "1881223807": 1.0,
129
+ "1919198188": 1.0,
130
+ "1991414747": 1.0,
131
+ "1997090716": 1.0,
132
+ "2015956957": 1.0,
133
+ "2054899126": 1.0,
134
+ "211245842": 1.0,
135
+ "2141802816": 1.0,
136
+ "256770874": 1.0,
137
+ "278008524": 1.0,
138
+ "416183106": 1.0,
139
+ "534717141": 1.0,
140
+ "551504635": 1.0,
141
+ "565369392": 1.0,
142
+ "565621826": 1.0,
143
+ "585904130": 1.0,
144
+ "626486813": 1.0,
145
+ "636680137": 1.0,
146
+ "706479600": 1.0,
147
+ "750448296": 1.0,
148
+ "75070641": 1.0,
149
+ "767747424": 1.0,
150
+ "77147264": 1.0,
151
+ "794181626": 1.0,
152
+ "957101489": 1.0,
153
+ "998761162": 1.0
154
+ }
155
+ },
156
+ "macro_by_task": {
157
+ "ci": {
158
+ "confidence": 0.95,
159
+ "high": 1.0,
160
+ "low": 1.0,
161
+ "mean": 1.0,
162
+ "n": 5,
163
+ "num_samples": 200
164
+ },
165
+ "mean": 1.0,
166
+ "per_task": {
167
+ "LiftPegUpright-v1": 1.0,
168
+ "PickCube-v1": 1.0,
169
+ "PullCube-v1": 1.0,
170
+ "PushCube-v1": 1.0,
171
+ "StackCube-v1": 1.0
172
+ }
173
+ },
174
+ "metric": "base_safety_label_known",
175
+ "micro": {
176
+ "confidence": 0.95,
177
+ "high": 1.0,
178
+ "low": 1.0,
179
+ "mean": 1.0,
180
+ "n": 69,
181
+ "num_samples": 200
182
+ },
183
+ "n": 69
184
+ },
185
+ "base_success": {
186
+ "macro_by_seed": {
187
+ "ci": {
188
+ "confidence": 0.95,
189
+ "high": 0.4273049645390071,
190
+ "low": 0.2024822695035461,
191
+ "mean": 0.32163120567375886,
192
+ "n": 47,
193
+ "num_samples": 200
194
+ },
195
+ "mean": 0.32163120567375886,
196
+ "per_seed": {
197
+ "1025740212": 0.0,
198
+ "1055157354": 0.0,
199
+ "1055567618": 1.0,
200
+ "1113268940": 0.0,
201
+ "1113371261": 1.0,
202
+ "1176943004": 1.0,
203
+ "1189392917": 1.0,
204
+ "1231780997": 0.0,
205
+ "1271718007": 0.5,
206
+ "1273247072": 0.0,
207
+ "1358421325": 1.0,
208
+ "1375492786": 0.0,
209
+ "141984615": 0.0,
210
+ "1475440973": 0.0,
211
+ "1483645336": 1.0,
212
+ "1522272686": 0.0,
213
+ "1528740417": 0.0,
214
+ "1549324805": 0.3333333333333333,
215
+ "1607773860": 0.0,
216
+ "1665739511": 1.0,
217
+ "1747178030": 1.0,
218
+ "1881223807": 0.0,
219
+ "1919198188": 0.0,
220
+ "1991414747": 1.0,
221
+ "1997090716": 1.0,
222
+ "2015956957": 0.2,
223
+ "2054899126": 1.0,
224
+ "211245842": 0.0,
225
+ "2141802816": 0.0,
226
+ "256770874": 0.0,
227
+ "278008524": 0.0,
228
+ "416183106": 0.0,
229
+ "534717141": 0.0,
230
+ "551504635": 0.0,
231
+ "565369392": 0.3333333333333333,
232
+ "565621826": 1.0,
233
+ "585904130": 0.0,
234
+ "626486813": 0.25,
235
+ "636680137": 0.0,
236
+ "706479600": 0.0,
237
+ "750448296": 0.0,
238
+ "75070641": 0.0,
239
+ "767747424": 0.0,
240
+ "77147264": 0.5,
241
+ "794181626": 1.0,
242
+ "957101489": 0.0,
243
+ "998761162": 0.0
244
+ }
245
+ },
246
+ "macro_by_task": {
247
+ "ci": {
248
+ "confidence": 0.95,
249
+ "high": 0.45656108597285067,
250
+ "low": 0.041176470588235294,
251
+ "mean": 0.21963800904977374,
252
+ "n": 5,
253
+ "num_samples": 200
254
+ },
255
+ "mean": 0.21963800904977374,
256
+ "per_task": {
257
+ "LiftPegUpright-v1": 0.0,
258
+ "PickCube-v1": 0.0,
259
+ "PullCube-v1": 0.20588235294117646,
260
+ "PushCube-v1": 0.6923076923076923,
261
+ "StackCube-v1": 0.2
262
+ }
263
+ },
264
+ "metric": "base_success",
265
+ "micro": {
266
+ "confidence": 0.95,
267
+ "high": 0.37681159420289856,
268
+ "low": 0.17391304347826086,
269
+ "mean": 0.2753623188405797,
270
+ "n": 69,
271
+ "num_samples": 200
272
+ },
273
+ "n": 69
274
+ },
275
+ "base_unsafe_known": {
276
+ "macro_by_seed": {
277
+ "ci": {
278
+ "confidence": 0.95,
279
+ "high": 0.0,
280
+ "low": 0.0,
281
+ "mean": 0.0,
282
+ "n": 47,
283
+ "num_samples": 200
284
+ },
285
+ "mean": 0.0,
286
+ "per_seed": {
287
+ "1025740212": 0.0,
288
+ "1055157354": 0.0,
289
+ "1055567618": 0.0,
290
+ "1113268940": 0.0,
291
+ "1113371261": 0.0,
292
+ "1176943004": 0.0,
293
+ "1189392917": 0.0,
294
+ "1231780997": 0.0,
295
+ "1271718007": 0.0,
296
+ "1273247072": 0.0,
297
+ "1358421325": 0.0,
298
+ "1375492786": 0.0,
299
+ "141984615": 0.0,
300
+ "1475440973": 0.0,
301
+ "1483645336": 0.0,
302
+ "1522272686": 0.0,
303
+ "1528740417": 0.0,
304
+ "1549324805": 0.0,
305
+ "1607773860": 0.0,
306
+ "1665739511": 0.0,
307
+ "1747178030": 0.0,
308
+ "1881223807": 0.0,
309
+ "1919198188": 0.0,
310
+ "1991414747": 0.0,
311
+ "1997090716": 0.0,
312
+ "2015956957": 0.0,
313
+ "2054899126": 0.0,
314
+ "211245842": 0.0,
315
+ "2141802816": 0.0,
316
+ "256770874": 0.0,
317
+ "278008524": 0.0,
318
+ "416183106": 0.0,
319
+ "534717141": 0.0,
320
+ "551504635": 0.0,
321
+ "565369392": 0.0,
322
+ "565621826": 0.0,
323
+ "585904130": 0.0,
324
+ "626486813": 0.0,
325
+ "636680137": 0.0,
326
+ "706479600": 0.0,
327
+ "750448296": 0.0,
328
+ "75070641": 0.0,
329
+ "767747424": 0.0,
330
+ "77147264": 0.0,
331
+ "794181626": 0.0,
332
+ "957101489": 0.0,
333
+ "998761162": 0.0
334
+ }
335
+ },
336
+ "macro_by_task": {
337
+ "ci": {
338
+ "confidence": 0.95,
339
+ "high": 0.0,
340
+ "low": 0.0,
341
+ "mean": 0.0,
342
+ "n": 5,
343
+ "num_samples": 200
344
+ },
345
+ "mean": 0.0,
346
+ "per_task": {
347
+ "LiftPegUpright-v1": 0.0,
348
+ "PickCube-v1": 0.0,
349
+ "PullCube-v1": 0.0,
350
+ "PushCube-v1": 0.0,
351
+ "StackCube-v1": 0.0
352
+ }
353
+ },
354
+ "metric": "base_unsafe_known",
355
+ "micro": {
356
+ "confidence": 0.95,
357
+ "high": 0.0,
358
+ "low": 0.0,
359
+ "mean": 0.0,
360
+ "n": 69,
361
+ "num_samples": 200
362
+ },
363
+ "n": 69
364
+ },
365
+ "base_utility": {
366
+ "macro_by_seed": {
367
+ "ci": {
368
+ "confidence": 0.95,
369
+ "high": 1.0092363624206038,
370
+ "low": 0.5980560999152297,
371
+ "mean": 0.827331273289437,
372
+ "n": 47,
373
+ "num_samples": 200
374
+ },
375
+ "mean": 0.827331273289437,
376
+ "per_seed": {
377
+ "1025740212": 0.21361251175403595,
378
+ "1055157354": 0.1879976987838745,
379
+ "1055567618": 2.0,
380
+ "1113268940": 0.21901588141918182,
381
+ "1113371261": 2.0,
382
+ "1176943004": 2.0,
383
+ "1189392917": 2.0,
384
+ "1231780997": 0.21144884824752808,
385
+ "1271718007": 1.10479174554348,
386
+ "1273247072": 0.11945180594921112,
387
+ "1358421325": 2.0,
388
+ "1375492786": 0.18323469161987305,
389
+ "141984615": 0.12483831991751988,
390
+ "1475440973": 0.7229747772216797,
391
+ "1483645336": 2.0,
392
+ "1522272686": 0.5529977083206177,
393
+ "1528740417": 0.32647061347961426,
394
+ "1549324805": 0.7632389987508456,
395
+ "1607773860": 0.8032726645469666,
396
+ "1665739511": 2.0,
397
+ "1747178030": 2.0,
398
+ "1881223807": 0.05987149477005005,
399
+ "1919198188": 0.06612768769264221,
400
+ "1991414747": 2.0,
401
+ "1997090716": 2.0,
402
+ "2015956957": 0.4729121312499046,
403
+ "2054899126": 2.0,
404
+ "211245842": 0.791279137134552,
405
+ "2141802816": 0.23664958775043488,
406
+ "256770874": 0.18160098791122437,
407
+ "278008524": 0.060655687004327774,
408
+ "416183106": 0.23001998662948608,
409
+ "534717141": 0.13276933133602142,
410
+ "551504635": 0.7272467017173767,
411
+ "565369392": 0.8445292512575785,
412
+ "565621826": 2.0,
413
+ "585904130": 0.13642114897569022,
414
+ "626486813": 0.5452363695949316,
415
+ "636680137": 0.07024133205413818,
416
+ "706479600": 0.8188905119895935,
417
+ "750448296": 0.14706942811608315,
418
+ "75070641": 0.194026380777359,
419
+ "767747424": 0.10894576646387577,
420
+ "77147264": 1.0903030335903168,
421
+ "794181626": 2.0,
422
+ "957101489": 0.23217551410198212,
423
+ "998761162": 0.20425210893154144
424
+ }
425
+ },
426
+ "macro_by_task": {
427
+ "ci": {
428
+ "confidence": 0.95,
429
+ "high": 1.03918900961902,
430
+ "low": 0.31294104572896864,
431
+ "mean": 0.6313663369707123,
432
+ "n": 5,
433
+ "num_samples": 200
434
+ },
435
+ "mean": 0.6313663369707123,
436
+ "per_task": {
437
+ "LiftPegUpright-v1": 0.40346155563990277,
438
+ "PickCube-v1": 0.11945180594921112,
439
+ "PullCube-v1": 0.5097779550534838,
440
+ "PushCube-v1": 1.4275685124672377,
441
+ "StackCube-v1": 0.696571855743726
442
+ }
443
+ },
444
+ "metric": "base_utility",
445
+ "micro": {
446
+ "confidence": 0.95,
447
+ "high": 0.895466687940601,
448
+ "low": 0.5355817867254001,
449
+ "mean": 0.7084005813019864,
450
+ "n": 69,
451
+ "num_samples": 200
452
+ },
453
+ "n": 69
454
+ },
455
+ "branch_car_at_8": {
456
+ "macro_by_seed": {
457
+ "ci": {
458
+ "confidence": 0.95,
459
+ "high": 0.2221361303896848,
460
+ "low": 0.03224338199580022,
461
+ "mean": 0.11875610934002392,
462
+ "n": 47,
463
+ "num_samples": 200
464
+ },
465
+ "mean": 0.11875610934002392,
466
+ "per_seed": {
467
+ "1025740212": 0.0009215474128723145,
468
+ "1055157354": 0.014723598957061768,
469
+ "1055567618": 0.018837809562683105,
470
+ "1113268940": 0.006647542119026184,
471
+ "1113371261": 0.0,
472
+ "1176943004": 0.03303612768650055,
473
+ "1189392917": 0.0,
474
+ "1231780997": 0.0,
475
+ "1271718007": 0.0,
476
+ "1273247072": 0.011674322187900543,
477
+ "1358421325": 0.6867138147354126,
478
+ "1375492786": 0.0017340581398457289,
479
+ "141984615": 0.017901106815164287,
480
+ "1475440973": 0.004936397075653076,
481
+ "1483645336": 1.8495931923389435,
482
+ "1522272686": 0.3386564999818802,
483
+ "1528740417": 0.02146562933921814,
484
+ "1549324805": 0.6191183490057787,
485
+ "1607773860": 0.011432766914367676,
486
+ "1665739511": 0.0,
487
+ "1747178030": 0.01012641191482544,
488
+ "1881223807": 0.0018659532070159912,
489
+ "1919198188": 0.0067754387855529785,
490
+ "1991414747": 0.004616528749465942,
491
+ "1997090716": 0.007344692945480347,
492
+ "2015956957": 0.01736527867615223,
493
+ "2054899126": 0.0,
494
+ "211245842": 0.04498988389968872,
495
+ "2141802816": 0.038818299770355225,
496
+ "256770874": 0.042443424463272095,
497
+ "278008524": 0.008101326413452625,
498
+ "416183106": 0.0,
499
+ "534717141": 0.0,
500
+ "551504635": 1.2730597853660583,
501
+ "565369392": 0.0,
502
+ "565621826": 0.0,
503
+ "585904130": 0.022781788061062496,
504
+ "626486813": 0.01009321678429842,
505
+ "636680137": 0.02178390882909298,
506
+ "706479600": 0.018285930156707764,
507
+ "750448296": 0.02600152976810932,
508
+ "75070641": 0.024178534746170044,
509
+ "767747424": 0.016496962402015924,
510
+ "77147264": 0.00890406221151352,
511
+ "794181626": 0.0,
512
+ "957101489": 0.3351810723543167,
513
+ "998761162": 0.004930347204208374
514
+ }
515
+ },
516
+ "macro_by_task": {
517
+ "ci": {
518
+ "confidence": 0.95,
519
+ "high": 0.16483040890711195,
520
+ "low": 0.04860531061887741,
521
+ "mean": 0.11001546229386132,
522
+ "n": 5,
523
+ "num_samples": 200
524
+ },
525
+ "mean": 0.11001546229386132,
526
+ "per_task": {
527
+ "LiftPegUpright-v1": 0.217838483552138,
528
+ "PickCube-v1": 0.011674322187900543,
529
+ "PullCube-v1": 0.06663480061380302,
530
+ "PushCube-v1": 0.1499279118501223,
531
+ "StackCube-v1": 0.10400179326534272
532
+ }
533
+ },
534
+ "metric": "branch_car_at_8",
535
+ "micro": {
536
+ "confidence": 0.95,
537
+ "high": 0.1787753005552551,
538
+ "low": 0.03932376045976644,
539
+ "mean": 0.10280258257104002,
540
+ "n": 69,
541
+ "num_samples": 200
542
+ },
543
+ "n": 69
544
+ },
545
+ "generated_safety_label_coverage_at_8": {
546
+ "macro_by_seed": {
547
+ "ci": {
548
+ "confidence": 0.95,
549
+ "high": 1.0,
550
+ "low": 1.0,
551
+ "mean": 1.0,
552
+ "n": 47,
553
+ "num_samples": 200
554
+ },
555
+ "mean": 1.0,
556
+ "per_seed": {
557
+ "1025740212": 1.0,
558
+ "1055157354": 1.0,
559
+ "1055567618": 1.0,
560
+ "1113268940": 1.0,
561
+ "1113371261": 1.0,
562
+ "1176943004": 1.0,
563
+ "1189392917": 1.0,
564
+ "1231780997": 1.0,
565
+ "1271718007": 1.0,
566
+ "1273247072": 1.0,
567
+ "1358421325": 1.0,
568
+ "1375492786": 1.0,
569
+ "141984615": 1.0,
570
+ "1475440973": 1.0,
571
+ "1483645336": 1.0,
572
+ "1522272686": 1.0,
573
+ "1528740417": 1.0,
574
+ "1549324805": 1.0,
575
+ "1607773860": 1.0,
576
+ "1665739511": 1.0,
577
+ "1747178030": 1.0,
578
+ "1881223807": 1.0,
579
+ "1919198188": 1.0,
580
+ "1991414747": 1.0,
581
+ "1997090716": 1.0,
582
+ "2015956957": 1.0,
583
+ "2054899126": 1.0,
584
+ "211245842": 1.0,
585
+ "2141802816": 1.0,
586
+ "256770874": 1.0,
587
+ "278008524": 1.0,
588
+ "416183106": 1.0,
589
+ "534717141": 1.0,
590
+ "551504635": 1.0,
591
+ "565369392": 1.0,
592
+ "565621826": 1.0,
593
+ "585904130": 1.0,
594
+ "626486813": 1.0,
595
+ "636680137": 1.0,
596
+ "706479600": 1.0,
597
+ "750448296": 1.0,
598
+ "75070641": 1.0,
599
+ "767747424": 1.0,
600
+ "77147264": 1.0,
601
+ "794181626": 1.0,
602
+ "957101489": 1.0,
603
+ "998761162": 1.0
604
+ }
605
+ },
606
+ "macro_by_task": {
607
+ "ci": {
608
+ "confidence": 0.95,
609
+ "high": 1.0,
610
+ "low": 1.0,
611
+ "mean": 1.0,
612
+ "n": 5,
613
+ "num_samples": 200
614
+ },
615
+ "mean": 1.0,
616
+ "per_task": {
617
+ "LiftPegUpright-v1": 1.0,
618
+ "PickCube-v1": 1.0,
619
+ "PullCube-v1": 1.0,
620
+ "PushCube-v1": 1.0,
621
+ "StackCube-v1": 1.0
622
+ }
623
+ },
624
+ "metric": "generated_safety_label_coverage_at_8",
625
+ "micro": {
626
+ "confidence": 0.95,
627
+ "high": 1.0,
628
+ "low": 1.0,
629
+ "mean": 1.0,
630
+ "n": 69,
631
+ "num_samples": 200
632
+ },
633
+ "n": 69
634
+ },
635
+ "generated_unsafe_rate_known_at_8": {
636
+ "macro_by_seed": {
637
+ "ci": {
638
+ "confidence": 0.95,
639
+ "high": 0.2677748226950355,
640
+ "low": 0.09539007092198581,
641
+ "mean": 0.18071808510638296,
642
+ "n": 47,
643
+ "num_samples": 200
644
+ },
645
+ "mean": 0.18071808510638296,
646
+ "per_seed": {
647
+ "1025740212": 0.0,
648
+ "1055157354": 0.0,
649
+ "1055567618": 0.0,
650
+ "1113268940": 0.0,
651
+ "1113371261": 0.0,
652
+ "1176943004": 0.0,
653
+ "1189392917": 0.0,
654
+ "1231780997": 0.0,
655
+ "1271718007": 0.0,
656
+ "1273247072": 0.0,
657
+ "1358421325": 0.0,
658
+ "1375492786": 1.0,
659
+ "141984615": 0.3333333333333333,
660
+ "1475440973": 0.0,
661
+ "1483645336": 0.0,
662
+ "1522272686": 0.0,
663
+ "1528740417": 0.0,
664
+ "1549324805": 0.75,
665
+ "1607773860": 0.0,
666
+ "1665739511": 0.0,
667
+ "1747178030": 1.0,
668
+ "1881223807": 0.0,
669
+ "1919198188": 0.0,
670
+ "1991414747": 0.0,
671
+ "1997090716": 0.0,
672
+ "2015956957": 0.525,
673
+ "2054899126": 1.0,
674
+ "211245842": 0.0,
675
+ "2141802816": 0.0,
676
+ "256770874": 0.0,
677
+ "278008524": 0.5,
678
+ "416183106": 0.0,
679
+ "534717141": 0.0,
680
+ "551504635": 0.0,
681
+ "565369392": 0.8333333333333334,
682
+ "565621826": 1.0,
683
+ "585904130": 0.3333333333333333,
684
+ "626486813": 0.28125,
685
+ "636680137": 0.25,
686
+ "706479600": 0.0,
687
+ "750448296": 0.1875,
688
+ "75070641": 0.0,
689
+ "767747424": 0.5,
690
+ "77147264": 0.0,
691
+ "794181626": 0.0,
692
+ "957101489": 0.0,
693
+ "998761162": 0.0
694
+ }
695
+ },
696
+ "macro_by_task": {
697
+ "ci": {
698
+ "confidence": 0.95,
699
+ "high": 0.31985294117647056,
700
+ "low": 0.0,
701
+ "mean": 0.10661764705882353,
702
+ "n": 5,
703
+ "num_samples": 200
704
+ },
705
+ "mean": 0.10661764705882353,
706
+ "per_task": {
707
+ "LiftPegUpright-v1": 0.0,
708
+ "PickCube-v1": 0.0,
709
+ "PullCube-v1": 0.5330882352941176,
710
+ "PushCube-v1": 0.0,
711
+ "StackCube-v1": 0.0
712
+ }
713
+ },
714
+ "metric": "generated_unsafe_rate_known_at_8",
715
+ "micro": {
716
+ "confidence": 0.95,
717
+ "high": 0.358695652173913,
718
+ "low": 0.17572463768115942,
719
+ "mean": 0.26268115942028986,
720
+ "n": 69,
721
+ "num_samples": 200
722
+ },
723
+ "n": 69
724
+ },
725
+ "hidden_chart_oracle_success_at_8": {
726
+ "macro_by_seed": {
727
+ "ci": {
728
+ "confidence": 0.95,
729
+ "high": 0.9148936170212765,
730
+ "low": 0.7432624113475178,
731
+ "mean": 0.8358156028368794,
732
+ "n": 47,
733
+ "num_samples": 200
734
+ },
735
+ "mean": 0.8358156028368794,
736
+ "per_seed": {
737
+ "1025740212": 1.0,
738
+ "1055157354": 1.0,
739
+ "1055567618": 1.0,
740
+ "1113268940": 1.0,
741
+ "1113371261": 1.0,
742
+ "1176943004": 1.0,
743
+ "1189392917": 1.0,
744
+ "1231780997": 1.0,
745
+ "1271718007": 1.0,
746
+ "1273247072": 1.0,
747
+ "1358421325": 1.0,
748
+ "1375492786": 0.0,
749
+ "141984615": 0.0,
750
+ "1475440973": 1.0,
751
+ "1483645336": 1.0,
752
+ "1522272686": 1.0,
753
+ "1528740417": 1.0,
754
+ "1549324805": 0.6666666666666666,
755
+ "1607773860": 1.0,
756
+ "1665739511": 1.0,
757
+ "1747178030": 1.0,
758
+ "1881223807": 1.0,
759
+ "1919198188": 1.0,
760
+ "1991414747": 1.0,
761
+ "1997090716": 1.0,
762
+ "2015956957": 0.2,
763
+ "2054899126": 1.0,
764
+ "211245842": 1.0,
765
+ "2141802816": 1.0,
766
+ "256770874": 1.0,
767
+ "278008524": 0.5,
768
+ "416183106": 1.0,
769
+ "534717141": 1.0,
770
+ "551504635": 1.0,
771
+ "565369392": 0.6666666666666666,
772
+ "565621826": 1.0,
773
+ "585904130": 0.0,
774
+ "626486813": 0.25,
775
+ "636680137": 1.0,
776
+ "706479600": 1.0,
777
+ "750448296": 0.0,
778
+ "75070641": 1.0,
779
+ "767747424": 0.0,
780
+ "77147264": 1.0,
781
+ "794181626": 1.0,
782
+ "957101489": 1.0,
783
+ "998761162": 1.0
784
+ }
785
+ },
786
+ "macro_by_task": {
787
+ "ci": {
788
+ "confidence": 0.95,
789
+ "high": 1.0,
790
+ "low": 0.5941176470588235,
791
+ "mean": 0.8647058823529413,
792
+ "n": 5,
793
+ "num_samples": 200
794
+ },
795
+ "mean": 0.8647058823529413,
796
+ "per_task": {
797
+ "LiftPegUpright-v1": 1.0,
798
+ "PickCube-v1": 1.0,
799
+ "PullCube-v1": 0.3235294117647059,
800
+ "PushCube-v1": 1.0,
801
+ "StackCube-v1": 1.0
802
+ }
803
+ },
804
+ "metric": "hidden_chart_oracle_success_at_8",
805
+ "micro": {
806
+ "confidence": 0.95,
807
+ "high": 0.782608695652174,
808
+ "low": 0.5507246376811594,
809
+ "mean": 0.6666666666666666,
810
+ "n": 69,
811
+ "num_samples": 200
812
+ },
813
+ "n": 69
814
+ },
815
+ "hidden_chart_oracle_utility_at_8": {
816
+ "macro_by_seed": {
817
+ "ci": {
818
+ "confidence": 0.95,
819
+ "high": 1.850041654515774,
820
+ "low": 1.5397213380149701,
821
+ "mean": 1.7062028111638965,
822
+ "n": 47,
823
+ "num_samples": 200
824
+ },
825
+ "mean": 1.7062028111638965,
826
+ "per_seed": {
827
+ "1025740212": 2.0,
828
+ "1055157354": 2.0,
829
+ "1055567618": 2.0,
830
+ "1113268940": 2.0,
831
+ "1113371261": 2.0,
832
+ "1176943004": 2.0,
833
+ "1189392917": 2.0,
834
+ "1231780997": 2.0,
835
+ "1271718007": 2.0,
836
+ "1273247072": 2.0,
837
+ "1358421325": 2.0,
838
+ "1375492786": 0.24345096945762634,
839
+ "141984615": 0.20044784247875214,
840
+ "1475440973": 2.0,
841
+ "1483645336": 2.0,
842
+ "1522272686": 2.0,
843
+ "1528740417": 2.0,
844
+ "1549324805": 1.4027505218982697,
845
+ "1607773860": 2.0,
846
+ "1665739511": 2.0,
847
+ "1747178030": 2.0,
848
+ "1881223807": 2.0,
849
+ "1919198188": 2.0,
850
+ "1991414747": 2.0,
851
+ "1997090716": 2.0,
852
+ "2015956957": 0.5298813670873642,
853
+ "2054899126": 2.0,
854
+ "211245842": 2.0,
855
+ "2141802816": 2.0,
856
+ "256770874": 2.0,
857
+ "278008524": 1.113811194896698,
858
+ "416183106": 2.0,
859
+ "534717141": 2.0,
860
+ "551504635": 2.0,
861
+ "565369392": 1.4266315797964733,
862
+ "565621826": 2.0,
863
+ "585904130": 0.23764699200789133,
864
+ "626486813": 0.6124684102833271,
865
+ "636680137": 2.0,
866
+ "706479600": 2.0,
867
+ "750448296": 0.23580759018659592,
868
+ "75070641": 2.0,
869
+ "767747424": 0.18863565661013126,
870
+ "77147264": 2.0,
871
+ "794181626": 2.0,
872
+ "957101489": 2.0,
873
+ "998761162": 2.0
874
+ }
875
+ },
876
+ "macro_by_task": {
877
+ "ci": {
878
+ "confidence": 0.95,
879
+ "high": 2.0,
880
+ "low": 1.2693931020796299,
881
+ "mean": 1.7564643673598765,
882
+ "n": 5,
883
+ "num_samples": 200
884
+ },
885
+ "mean": 1.7564643673598765,
886
+ "per_task": {
887
+ "LiftPegUpright-v1": 2.0,
888
+ "PickCube-v1": 2.0,
889
+ "PullCube-v1": 0.7823218367993832,
890
+ "PushCube-v1": 2.0,
891
+ "StackCube-v1": 2.0
892
+ }
893
+ },
894
+ "metric": "hidden_chart_oracle_utility_at_8",
895
+ "micro": {
896
+ "confidence": 0.95,
897
+ "high": 1.6080652458080347,
898
+ "low": 1.1926357444861662,
899
+ "mean": 1.3999846732054932,
900
+ "n": 69,
901
+ "num_samples": 200
902
+ },
903
+ "n": 69
904
+ },
905
+ "ncar_to_hidden_chart_oracle_at_8": {
906
+ "macro_by_seed": {
907
+ "ci": {
908
+ "confidence": 0.95,
909
+ "high": 1.9732685111326405,
910
+ "low": 1.2193770303676617,
911
+ "mean": 1.5828590785720633,
912
+ "n": 34,
913
+ "num_samples": 200
914
+ },
915
+ "mean": 1.5828590785720633,
916
+ "per_seed": {
917
+ "1025740212": 1.0038598175480926,
918
+ "1055157354": 0.9915817866344643,
919
+ "1113268940": 1.0031982588625619,
920
+ "1231780997": 1.0005967986281998,
921
+ "1271718007": 0.977550125740515,
922
+ "1273247072": 1.0157636165911068,
923
+ "1375492786": 3.9862911723153553,
924
+ "141984615": 3.2387927981827365,
925
+ "1475440973": 0.9997688595390668,
926
+ "1522272686": 1.2270471627342485,
927
+ "1528740417": 1.0390089371449063,
928
+ "1549324805": 2.43115209329291,
929
+ "1607773860": 1.0017699108028124,
930
+ "1881223807": 1.0055300485453837,
931
+ "1919198188": 1.0050267795980445,
932
+ "2015956957": 1.4284452600977113,
933
+ "211245842": 1.0007633933743367,
934
+ "2141802816": 1.0465578663511705,
935
+ "256770874": 1.002821760012782,
936
+ "278008524": 1.1481789711765635,
937
+ "416183106": 0.828016730744017,
938
+ "534717141": 1.0230647928623209,
939
+ "551504635": 1.0002407984968027,
940
+ "565369392": 4.740664659244946,
941
+ "585904130": 1.5916216907065317,
942
+ "626486813": 0.8102411086604181,
943
+ "636680137": 1.0226806271202884,
944
+ "706479600": 1.0001120742118061,
945
+ "750448296": 4.978844987783829,
946
+ "75070641": 1.0008596290963836,
947
+ "767747424": 5.205313661372345,
948
+ "77147264": 1.0010257406519596,
949
+ "957101489": 0.998510000020915,
950
+ "998761162": 1.062306753304616
951
+ }
952
+ },
953
+ "macro_by_task": {
954
+ "ci": {
955
+ "confidence": 0.95,
956
+ "high": 2.075213630454681,
957
+ "low": 1.0074337911202729,
958
+ "mean": 1.3676944672031914,
959
+ "n": 5,
960
+ "num_samples": 200
961
+ },
962
+ "mean": 1.3676944672031914,
963
+ "per_task": {
964
+ "LiftPegUpright-v1": 1.0077789117032718,
965
+ "PickCube-v1": 1.0157636165911068,
966
+ "PullCube-v1": 2.7805691722759143,
967
+ "PushCube-v1": 1.0314368776438072,
968
+ "StackCube-v1": 1.002923757801857
969
+ }
970
+ },
971
+ "metric": "ncar_to_hidden_chart_oracle_at_8",
972
+ "micro": {
973
+ "confidence": 0.95,
974
+ "high": 2.9682945900917206,
975
+ "low": 1.3091031892701088,
976
+ "mean": 1.9659727468491581,
977
+ "n": 50,
978
+ "num_samples": 200
979
+ },
980
+ "n": 50
981
+ },
982
+ "ncar_to_proposal_oracle_at_8": {
983
+ "macro_by_seed": {
984
+ "ci": {
985
+ "confidence": 0.95,
986
+ "high": 3.2311590033878983,
987
+ "low": 0.46294720224637953,
988
+ "mean": 1.4953472093017477,
989
+ "n": 39,
990
+ "num_samples": 200
991
+ },
992
+ "mean": 1.4953472093017477,
993
+ "per_seed": {
994
+ "1025740212": 0.15426996158755663,
995
+ "1055157354": 0.49115629197991395,
996
+ "1055567618": 0.016240555308914228,
997
+ "1176943004": 0.01735576759836683,
998
+ "1231780997": 0.0,
999
+ "1271718007": 0.0,
1000
+ "1273247072": 0.649658432039702,
1001
+ "1358421325": 0.5920648030985759,
1002
+ "1375492786": 0.009737013067671323,
1003
+ "141984615": 0.20565346679569363,
1004
+ "1475440973": 0.943578650052657,
1005
+ "1522272686": 33.46827145305134,
1006
+ "1528740417": 0.48989279105464534,
1007
+ "1549324805": 0.21532264167370982,
1008
+ "1607773860": 1.2273941543534532,
1009
+ "1747178030": 0.0052936433657889735,
1010
+ "1881223807": 0.21053121053631743,
1011
+ "1919198188": 2.3000872930411496,
1012
+ "1991414747": 0.002410605043679449,
1013
+ "1997090716": 0.003931374273492085,
1014
+ "2015956957": 0.3825324409775967,
1015
+ "211245842": 1.0209391670158559,
1016
+ "2141802816": 0.8969200607232635,
1017
+ "256770874": 1.1375171273285567,
1018
+ "278008524": 0.4289430850076562,
1019
+ "416183106": 0.0,
1020
+ "534717141": 0.0,
1021
+ "551504635": 1.0002407984968027,
1022
+ "565369392": 0.0,
1023
+ "585904130": 0.5193742667883882,
1024
+ "626486813": 0.2258722711825077,
1025
+ "636680137": 0.9908871412464478,
1026
+ "706479600": 1.0072917889865605,
1027
+ "750448296": 5.576871543359603,
1028
+ "75070641": 1.0686141104703737,
1029
+ "767747424": 0.9043237597220475,
1030
+ "77147264": 1.1170642490674998,
1031
+ "957101489": 0.9922026628719985,
1032
+ "998761162": 0.04609658160039006
1033
+ }
1034
+ },
1035
+ "macro_by_task": {
1036
+ "ci": {
1037
+ "confidence": 0.95,
1038
+ "high": 2.1379728454516433,
1039
+ "low": 0.4183144702771172,
1040
+ "mean": 1.0590787682193088,
1041
+ "n": 5,
1042
+ "num_samples": 200
1043
+ },
1044
+ "mean": 1.0590787682193088,
1045
+ "per_task": {
1046
+ "LiftPegUpright-v1": 0.5175964402383324,
1047
+ "PickCube-v1": 0.649658432039702,
1048
+ "PullCube-v1": 0.6898506880563886,
1049
+ "PushCube-v1": 0.26408516243539404,
1050
+ "StackCube-v1": 3.1742031183267274
1051
+ }
1052
+ },
1053
+ "metric": "ncar_to_proposal_oracle_at_8",
1054
+ "micro": {
1055
+ "confidence": 0.95,
1056
+ "high": 2.6377697726073546,
1057
+ "low": 0.44403409132042027,
1058
+ "mean": 1.2225939958034715,
1059
+ "n": 58,
1060
+ "num_samples": 200
1061
+ },
1062
+ "n": 58
1063
+ },
1064
+ "outcome_ptr_at_8": {
1065
+ "macro_by_seed": {
1066
+ "ci": {
1067
+ "confidence": 0.95,
1068
+ "high": 0.46276595744680843,
1069
+ "low": 0.24468085106382978,
1070
+ "mean": 0.3329787234042554,
1071
+ "n": 47,
1072
+ "num_samples": 200
1073
+ },
1074
+ "mean": 0.3329787234042554,
1075
+ "per_seed": {
1076
+ "1025740212": 0.0,
1077
+ "1055157354": 1.0,
1078
+ "1055567618": 0.0,
1079
+ "1113268940": 1.0,
1080
+ "1113371261": 0.0,
1081
+ "1176943004": 0.0,
1082
+ "1189392917": 0.0,
1083
+ "1231780997": 0.0,
1084
+ "1271718007": 0.5,
1085
+ "1273247072": 0.0,
1086
+ "1358421325": 0.0,
1087
+ "1375492786": 0.0,
1088
+ "141984615": 0.3333333333333333,
1089
+ "1475440973": 1.0,
1090
+ "1483645336": 0.0,
1091
+ "1522272686": 1.0,
1092
+ "1528740417": 0.0,
1093
+ "1549324805": 0.0,
1094
+ "1607773860": 1.0,
1095
+ "1665739511": 0.0,
1096
+ "1747178030": 0.0,
1097
+ "1881223807": 0.0,
1098
+ "1919198188": 0.0,
1099
+ "1991414747": 0.0,
1100
+ "1997090716": 0.0,
1101
+ "2015956957": 0.4,
1102
+ "2054899126": 0.0,
1103
+ "211245842": 1.0,
1104
+ "2141802816": 0.0,
1105
+ "256770874": 1.0,
1106
+ "278008524": 0.0,
1107
+ "416183106": 1.0,
1108
+ "534717141": 0.0,
1109
+ "551504635": 1.0,
1110
+ "565369392": 0.3333333333333333,
1111
+ "565621826": 0.0,
1112
+ "585904130": 0.3333333333333333,
1113
+ "626486813": 0.25,
1114
+ "636680137": 0.0,
1115
+ "706479600": 1.0,
1116
+ "750448296": 0.5,
1117
+ "75070641": 1.0,
1118
+ "767747424": 0.5,
1119
+ "77147264": 0.5,
1120
+ "794181626": 0.0,
1121
+ "957101489": 1.0,
1122
+ "998761162": 0.0
1123
+ }
1124
+ },
1125
+ "macro_by_task": {
1126
+ "ci": {
1127
+ "confidence": 0.95,
1128
+ "high": 0.5929411764705882,
1129
+ "low": 0.1,
1130
+ "mean": 0.2996078431372549,
1131
+ "n": 5,
1132
+ "num_samples": 200
1133
+ },
1134
+ "mean": 0.2996078431372549,
1135
+ "per_task": {
1136
+ "LiftPegUpright-v1": 0.5,
1137
+ "PickCube-v1": 0.0,
1138
+ "PullCube-v1": 0.2647058823529412,
1139
+ "PushCube-v1": 0.0,
1140
+ "StackCube-v1": 0.7333333333333333
1141
+ }
1142
+ },
1143
+ "metric": "outcome_ptr_at_8",
1144
+ "micro": {
1145
+ "confidence": 0.95,
1146
+ "high": 0.43478260869565216,
1147
+ "low": 0.21739130434782608,
1148
+ "mean": 0.3333333333333333,
1149
+ "n": 69,
1150
+ "num_samples": 200
1151
+ },
1152
+ "n": 69
1153
+ },
1154
+ "pairwise_causal_calibration_ece": {
1155
+ "macro_by_seed": {
1156
+ "ci": {
1157
+ "confidence": 0.95,
1158
+ "high": 0.24434323059457133,
1159
+ "low": 0.16350127765448375,
1160
+ "mean": 0.20129853329177672,
1161
+ "n": 47,
1162
+ "num_samples": 200
1163
+ },
1164
+ "mean": 0.20129853329177672,
1165
+ "per_seed": {
1166
+ "1025740212": 0.23026774639555236,
1167
+ "1055157354": 0.19960528924679433,
1168
+ "1055567618": 0.18844312444633177,
1169
+ "1113268940": 0.4201805991515518,
1170
+ "1113371261": 0.10528846281262146,
1171
+ "1176943004": 0.18726924478108561,
1172
+ "1189392917": 0.0,
1173
+ "1231780997": 0.13464701141472765,
1174
+ "1271718007": 0.07020098827281387,
1175
+ "1273247072": 0.11483608299984571,
1176
+ "1358421325": 0.24959422175257154,
1177
+ "1375492786": 0.13157026218508128,
1178
+ "141984615": 0.17017158124656792,
1179
+ "1475440973": 0.11407560353818716,
1180
+ "1483645336": 0.5238458588992289,
1181
+ "1522272686": 0.0746614940687875,
1182
+ "1528740417": 0.13131517178411375,
1183
+ "1549324805": 0.3542532249847814,
1184
+ "1607773860": 0.251291053201395,
1185
+ "1665739511": 0.0,
1186
+ "1747178030": 0.16974044575049851,
1187
+ "1881223807": 0.07419200542238083,
1188
+ "1919198188": 0.16604478136348194,
1189
+ "1991414747": 0.3416902568888008,
1190
+ "1997090716": 0.300657409601606,
1191
+ "2015956957": 0.24286396292452275,
1192
+ "2054899126": 0.0,
1193
+ "211245842": 0.4141064151265617,
1194
+ "2141802816": 0.3985289040635325,
1195
+ "256770874": 0.26681470238264204,
1196
+ "278008524": 0.20536892413092878,
1197
+ "416183106": 0.07874347996488716,
1198
+ "534717141": 0.0753952117725891,
1199
+ "551504635": 0.2344395807565722,
1200
+ "565369392": 0.13140272637925274,
1201
+ "565621826": 0.0,
1202
+ "585904130": 0.17283993917588428,
1203
+ "626486813": 0.2769376371795747,
1204
+ "636680137": 0.5932483433892377,
1205
+ "706479600": 0.28238485246274814,
1206
+ "750448296": 0.19889074905685378,
1207
+ "75070641": 0.1631929231601434,
1208
+ "767747424": 0.24962611487178488,
1209
+ "77147264": 0.05996298782788377,
1210
+ "794181626": 0.0,
1211
+ "957101489": 0.4692653473226291,
1212
+ "998761162": 0.24317634255647042
1213
+ }
1214
+ },
1215
+ "macro_by_task": {
1216
+ "ci": {
1217
+ "confidence": 0.95,
1218
+ "high": 0.21222831135604014,
1219
+ "low": 0.15062100011191726,
1220
+ "mean": 0.18656867805065086,
1221
+ "n": 5,
1222
+ "num_samples": 200
1223
+ },
1224
+ "mean": 0.18656867805065086,
1225
+ "per_task": {
1226
+ "LiftPegUpright-v1": 0.2108209521734508,
1227
+ "PickCube-v1": 0.11483608299984571,
1228
+ "PullCube-v1": 0.21534074294224181,
1229
+ "PushCube-v1": 0.1875472363576914,
1230
+ "StackCube-v1": 0.20429837578002458
1231
+ }
1232
+ },
1233
+ "metric": "pairwise_causal_calibration_ece",
1234
+ "micro": {
1235
+ "confidence": 0.95,
1236
+ "high": 0.23561359353056474,
1237
+ "low": 0.16483538000162765,
1238
+ "mean": 0.205854156020683,
1239
+ "n": 69,
1240
+ "num_samples": 200
1241
+ },
1242
+ "n": 69
1243
+ },
1244
+ "proposal_oracle_safety_label_known_at_8": {
1245
+ "macro_by_seed": {
1246
+ "ci": {
1247
+ "confidence": 0.95,
1248
+ "high": 1.0,
1249
+ "low": 1.0,
1250
+ "mean": 1.0,
1251
+ "n": 47,
1252
+ "num_samples": 200
1253
+ },
1254
+ "mean": 1.0,
1255
+ "per_seed": {
1256
+ "1025740212": 1.0,
1257
+ "1055157354": 1.0,
1258
+ "1055567618": 1.0,
1259
+ "1113268940": 1.0,
1260
+ "1113371261": 1.0,
1261
+ "1176943004": 1.0,
1262
+ "1189392917": 1.0,
1263
+ "1231780997": 1.0,
1264
+ "1271718007": 1.0,
1265
+ "1273247072": 1.0,
1266
+ "1358421325": 1.0,
1267
+ "1375492786": 1.0,
1268
+ "141984615": 1.0,
1269
+ "1475440973": 1.0,
1270
+ "1483645336": 1.0,
1271
+ "1522272686": 1.0,
1272
+ "1528740417": 1.0,
1273
+ "1549324805": 1.0,
1274
+ "1607773860": 1.0,
1275
+ "1665739511": 1.0,
1276
+ "1747178030": 1.0,
1277
+ "1881223807": 1.0,
1278
+ "1919198188": 1.0,
1279
+ "1991414747": 1.0,
1280
+ "1997090716": 1.0,
1281
+ "2015956957": 1.0,
1282
+ "2054899126": 1.0,
1283
+ "211245842": 1.0,
1284
+ "2141802816": 1.0,
1285
+ "256770874": 1.0,
1286
+ "278008524": 1.0,
1287
+ "416183106": 1.0,
1288
+ "534717141": 1.0,
1289
+ "551504635": 1.0,
1290
+ "565369392": 1.0,
1291
+ "565621826": 1.0,
1292
+ "585904130": 1.0,
1293
+ "626486813": 1.0,
1294
+ "636680137": 1.0,
1295
+ "706479600": 1.0,
1296
+ "750448296": 1.0,
1297
+ "75070641": 1.0,
1298
+ "767747424": 1.0,
1299
+ "77147264": 1.0,
1300
+ "794181626": 1.0,
1301
+ "957101489": 1.0,
1302
+ "998761162": 1.0
1303
+ }
1304
+ },
1305
+ "macro_by_task": {
1306
+ "ci": {
1307
+ "confidence": 0.95,
1308
+ "high": 1.0,
1309
+ "low": 1.0,
1310
+ "mean": 1.0,
1311
+ "n": 5,
1312
+ "num_samples": 200
1313
+ },
1314
+ "mean": 1.0,
1315
+ "per_task": {
1316
+ "LiftPegUpright-v1": 1.0,
1317
+ "PickCube-v1": 1.0,
1318
+ "PullCube-v1": 1.0,
1319
+ "PushCube-v1": 1.0,
1320
+ "StackCube-v1": 1.0
1321
+ }
1322
+ },
1323
+ "metric": "proposal_oracle_safety_label_known_at_8",
1324
+ "micro": {
1325
+ "confidence": 0.95,
1326
+ "high": 1.0,
1327
+ "low": 1.0,
1328
+ "mean": 1.0,
1329
+ "n": 69,
1330
+ "num_samples": 200
1331
+ },
1332
+ "n": 69
1333
+ },
1334
+ "proposal_oracle_success_at_8": {
1335
+ "macro_by_seed": {
1336
+ "ci": {
1337
+ "confidence": 0.95,
1338
+ "high": 0.3049645390070922,
1339
+ "low": 0.09574468085106383,
1340
+ "mean": 0.19858156028368795,
1341
+ "n": 47,
1342
+ "num_samples": 200
1343
+ },
1344
+ "mean": 0.19858156028368795,
1345
+ "per_seed": {
1346
+ "1025740212": 0.0,
1347
+ "1055157354": 0.0,
1348
+ "1055567618": 0.0,
1349
+ "1113268940": 0.0,
1350
+ "1113371261": 1.0,
1351
+ "1176943004": 0.0,
1352
+ "1189392917": 1.0,
1353
+ "1231780997": 0.0,
1354
+ "1271718007": 0.5,
1355
+ "1273247072": 0.0,
1356
+ "1358421325": 0.0,
1357
+ "1375492786": 0.0,
1358
+ "141984615": 0.0,
1359
+ "1475440973": 0.0,
1360
+ "1483645336": 1.0,
1361
+ "1522272686": 0.0,
1362
+ "1528740417": 0.0,
1363
+ "1549324805": 0.3333333333333333,
1364
+ "1607773860": 0.0,
1365
+ "1665739511": 1.0,
1366
+ "1747178030": 0.0,
1367
+ "1881223807": 0.0,
1368
+ "1919198188": 0.0,
1369
+ "1991414747": 0.0,
1370
+ "1997090716": 0.0,
1371
+ "2015956957": 0.0,
1372
+ "2054899126": 1.0,
1373
+ "211245842": 0.0,
1374
+ "2141802816": 0.0,
1375
+ "256770874": 0.0,
1376
+ "278008524": 0.0,
1377
+ "416183106": 0.0,
1378
+ "534717141": 0.0,
1379
+ "551504635": 1.0,
1380
+ "565369392": 0.0,
1381
+ "565621826": 1.0,
1382
+ "585904130": 0.0,
1383
+ "626486813": 0.0,
1384
+ "636680137": 0.0,
1385
+ "706479600": 0.0,
1386
+ "750448296": 0.0,
1387
+ "75070641": 0.0,
1388
+ "767747424": 0.0,
1389
+ "77147264": 0.5,
1390
+ "794181626": 1.0,
1391
+ "957101489": 0.0,
1392
+ "998761162": 0.0
1393
+ }
1394
+ },
1395
+ "macro_by_task": {
1396
+ "ci": {
1397
+ "confidence": 0.95,
1398
+ "high": 0.3235897435897436,
1399
+ "low": 0.03529411764705882,
1400
+ "mean": 0.15662141779788838,
1401
+ "n": 5,
1402
+ "num_samples": 200
1403
+ },
1404
+ "mean": 0.15662141779788838,
1405
+ "per_task": {
1406
+ "LiftPegUpright-v1": 0.16666666666666666,
1407
+ "PickCube-v1": 0.0,
1408
+ "PullCube-v1": 0.08823529411764706,
1409
+ "PushCube-v1": 0.46153846153846156,
1410
+ "StackCube-v1": 0.06666666666666667
1411
+ }
1412
+ },
1413
+ "metric": "proposal_oracle_success_at_8",
1414
+ "micro": {
1415
+ "confidence": 0.95,
1416
+ "high": 0.2608695652173913,
1417
+ "low": 0.08695652173913043,
1418
+ "mean": 0.15942028985507245,
1419
+ "n": 69,
1420
+ "num_samples": 200
1421
+ },
1422
+ "n": 69
1423
+ },
1424
+ "proposal_oracle_success_gain_over_base_at_8": {
1425
+ "macro_by_seed": {
1426
+ "ci": {
1427
+ "confidence": 0.95,
1428
+ "high": -0.018085106382978725,
1429
+ "low": -0.23120567375886522,
1430
+ "mean": -0.12304964539007092,
1431
+ "n": 47,
1432
+ "num_samples": 200
1433
+ },
1434
+ "mean": -0.12304964539007092,
1435
+ "per_seed": {
1436
+ "1025740212": 0.0,
1437
+ "1055157354": 0.0,
1438
+ "1055567618": -1.0,
1439
+ "1113268940": 0.0,
1440
+ "1113371261": 0.0,
1441
+ "1176943004": -1.0,
1442
+ "1189392917": 0.0,
1443
+ "1231780997": 0.0,
1444
+ "1271718007": 0.0,
1445
+ "1273247072": 0.0,
1446
+ "1358421325": -1.0,
1447
+ "1375492786": 0.0,
1448
+ "141984615": 0.0,
1449
+ "1475440973": 0.0,
1450
+ "1483645336": 0.0,
1451
+ "1522272686": 0.0,
1452
+ "1528740417": 0.0,
1453
+ "1549324805": 0.0,
1454
+ "1607773860": 0.0,
1455
+ "1665739511": 0.0,
1456
+ "1747178030": -1.0,
1457
+ "1881223807": 0.0,
1458
+ "1919198188": 0.0,
1459
+ "1991414747": -1.0,
1460
+ "1997090716": -1.0,
1461
+ "2015956957": -0.2,
1462
+ "2054899126": 0.0,
1463
+ "211245842": 0.0,
1464
+ "2141802816": 0.0,
1465
+ "256770874": 0.0,
1466
+ "278008524": 0.0,
1467
+ "416183106": 0.0,
1468
+ "534717141": 0.0,
1469
+ "551504635": 1.0,
1470
+ "565369392": -0.3333333333333333,
1471
+ "565621826": 0.0,
1472
+ "585904130": 0.0,
1473
+ "626486813": -0.25,
1474
+ "636680137": 0.0,
1475
+ "706479600": 0.0,
1476
+ "750448296": 0.0,
1477
+ "75070641": 0.0,
1478
+ "767747424": 0.0,
1479
+ "77147264": 0.0,
1480
+ "794181626": 0.0,
1481
+ "957101489": 0.0,
1482
+ "998761162": 0.0
1483
+ }
1484
+ },
1485
+ "macro_by_task": {
1486
+ "ci": {
1487
+ "confidence": 0.95,
1488
+ "high": 0.05294117647058823,
1489
+ "low": -0.16603318250377075,
1490
+ "mean": -0.06301659125188537,
1491
+ "n": 5,
1492
+ "num_samples": 200
1493
+ },
1494
+ "mean": -0.06301659125188537,
1495
+ "per_task": {
1496
+ "LiftPegUpright-v1": 0.16666666666666666,
1497
+ "PickCube-v1": 0.0,
1498
+ "PullCube-v1": -0.11764705882352941,
1499
+ "PushCube-v1": -0.23076923076923078,
1500
+ "StackCube-v1": -0.13333333333333333
1501
+ }
1502
+ },
1503
+ "metric": "proposal_oracle_success_gain_over_base_at_8",
1504
+ "micro": {
1505
+ "confidence": 0.95,
1506
+ "high": -0.043478260869565216,
1507
+ "low": -0.18840579710144928,
1508
+ "mean": -0.11594202898550725,
1509
+ "n": 69,
1510
+ "num_samples": 200
1511
+ },
1512
+ "n": 69
1513
+ },
1514
+ "proposal_oracle_unsafe_known_at_8": {
1515
+ "macro_by_seed": {
1516
+ "ci": {
1517
+ "confidence": 0.95,
1518
+ "high": 0.2457446808510638,
1519
+ "low": 0.08368794326241136,
1520
+ "mean": 0.16702127659574467,
1521
+ "n": 47,
1522
+ "num_samples": 200
1523
+ },
1524
+ "mean": 0.16702127659574467,
1525
+ "per_seed": {
1526
+ "1025740212": 0.0,
1527
+ "1055157354": 0.0,
1528
+ "1055567618": 0.0,
1529
+ "1113268940": 0.0,
1530
+ "1113371261": 0.0,
1531
+ "1176943004": 0.0,
1532
+ "1189392917": 0.0,
1533
+ "1231780997": 0.0,
1534
+ "1271718007": 0.0,
1535
+ "1273247072": 0.0,
1536
+ "1358421325": 0.0,
1537
+ "1375492786": 1.0,
1538
+ "141984615": 0.3333333333333333,
1539
+ "1475440973": 0.0,
1540
+ "1483645336": 0.0,
1541
+ "1522272686": 0.0,
1542
+ "1528740417": 0.0,
1543
+ "1549324805": 0.6666666666666666,
1544
+ "1607773860": 0.0,
1545
+ "1665739511": 0.0,
1546
+ "1747178030": 1.0,
1547
+ "1881223807": 0.0,
1548
+ "1919198188": 0.0,
1549
+ "1991414747": 0.0,
1550
+ "1997090716": 0.0,
1551
+ "2015956957": 0.6,
1552
+ "2054899126": 1.0,
1553
+ "211245842": 0.0,
1554
+ "2141802816": 0.0,
1555
+ "256770874": 0.0,
1556
+ "278008524": 0.5,
1557
+ "416183106": 0.0,
1558
+ "534717141": 0.0,
1559
+ "551504635": 0.0,
1560
+ "565369392": 0.6666666666666666,
1561
+ "565621826": 1.0,
1562
+ "585904130": 0.3333333333333333,
1563
+ "626486813": 0.25,
1564
+ "636680137": 0.0,
1565
+ "706479600": 0.0,
1566
+ "750448296": 0.0,
1567
+ "75070641": 0.0,
1568
+ "767747424": 0.5,
1569
+ "77147264": 0.0,
1570
+ "794181626": 0.0,
1571
+ "957101489": 0.0,
1572
+ "998761162": 0.0
1573
+ }
1574
+ },
1575
+ "macro_by_task": {
1576
+ "ci": {
1577
+ "confidence": 0.95,
1578
+ "high": 0.3,
1579
+ "low": 0.0,
1580
+ "mean": 0.1,
1581
+ "n": 5,
1582
+ "num_samples": 200
1583
+ },
1584
+ "mean": 0.1,
1585
+ "per_task": {
1586
+ "LiftPegUpright-v1": 0.0,
1587
+ "PickCube-v1": 0.0,
1588
+ "PullCube-v1": 0.5,
1589
+ "PushCube-v1": 0.0,
1590
+ "StackCube-v1": 0.0
1591
+ }
1592
+ },
1593
+ "metric": "proposal_oracle_unsafe_known_at_8",
1594
+ "micro": {
1595
+ "confidence": 0.95,
1596
+ "high": 0.34782608695652173,
1597
+ "low": 0.17391304347826086,
1598
+ "mean": 0.2463768115942029,
1599
+ "n": 69,
1600
+ "num_samples": 200
1601
+ },
1602
+ "n": 69
1603
+ },
1604
+ "proposal_oracle_utility_at_8": {
1605
+ "macro_by_seed": {
1606
+ "ci": {
1607
+ "confidence": 0.95,
1608
+ "high": 0.8188112276943124,
1609
+ "low": 0.42695560427316537,
1610
+ "mean": 0.6175774899780327,
1611
+ "n": 47,
1612
+ "num_samples": 200
1613
+ },
1614
+ "mean": 0.6175774899780327,
1615
+ "per_seed": {
1616
+ "1025740212": 0.2076389193534851,
1617
+ "1055157354": 0.21797510981559753,
1618
+ "1055567618": 0.8400760293006897,
1619
+ "1113268940": 0.21996736526489258,
1620
+ "1113371261": 2.0,
1621
+ "1176943004": 0.09653319418430328,
1622
+ "1189392917": 2.0,
1623
+ "1231780997": 0.2103814333677292,
1624
+ "1271718007": 1.1248890534043312,
1625
+ "1273247072": 0.10148187726736069,
1626
+ "1358421325": 0.8401374220848083,
1627
+ "1375492786": 0.005145370960235596,
1628
+ "141984615": 0.10611975689729054,
1629
+ "1475440973": 0.7282063364982605,
1630
+ "1483645336": 2.0,
1631
+ "1522272686": 0.5631164312362671,
1632
+ "1528740417": 0.2826536297798157,
1633
+ "1549324805": 0.7099594213068485,
1634
+ "1607773860": 0.812587320804596,
1635
+ "1665739511": 2.0,
1636
+ "1747178030": 0.08706194162368774,
1637
+ "1881223807": 0.0510084331035614,
1638
+ "1919198188": 0.06318196654319763,
1639
+ "1991414747": 0.08490870893001556,
1640
+ "1997090716": 0.1317746937274933,
1641
+ "2015956957": 0.11907548382878304,
1642
+ "2054899126": 2.0,
1643
+ "211245842": 0.8353462815284729,
1644
+ "2141802816": 0.1933700442314148,
1645
+ "256770874": 0.21891331672668457,
1646
+ "278008524": 0.03437099326401949,
1647
+ "416183106": 0.5344269275665283,
1648
+ "534717141": 0.08970203250646591,
1649
+ "551504635": 2.0,
1650
+ "565369392": 0.20857956757148108,
1651
+ "565621826": 2.0,
1652
+ "585904130": 0.09142490103840828,
1653
+ "626486813": 0.09146204218268394,
1654
+ "636680137": 0.048257093876600266,
1655
+ "706479600": 0.8370440602302551,
1656
+ "750448296": 0.14822042733430862,
1657
+ "75070641": 0.21665243804454803,
1658
+ "767747424": 0.10893216636031866,
1659
+ "77147264": 1.098273977637291,
1660
+ "794181626": 2.0,
1661
+ "957101489": 0.5699906349182129,
1662
+ "998761162": 0.09729522466659546
1663
+ }
1664
+ },
1665
+ "macro_by_task": {
1666
+ "ci": {
1667
+ "confidence": 0.95,
1668
+ "high": 0.7624335592781188,
1669
+ "low": 0.27158725250731497,
1670
+ "mean": 0.5119254678054566,
1671
+ "n": 5,
1672
+ "num_samples": 200
1673
+ },
1674
+ "mean": 0.5119254678054566,
1675
+ "per_task": {
1676
+ "LiftPegUpright-v1": 0.6081412807106972,
1677
+ "PickCube-v1": 0.10148187726736069,
1678
+ "PullCube-v1": 0.27341561364557815,
1679
+ "PushCube-v1": 0.9768703006781064,
1680
+ "StackCube-v1": 0.5997182667255402
1681
+ }
1682
+ },
1683
+ "metric": "proposal_oracle_utility_at_8",
1684
+ "micro": {
1685
+ "confidence": 0.95,
1686
+ "high": 0.6940592441005983,
1687
+ "low": 0.35817475390175113,
1688
+ "mean": 0.5035007005098505,
1689
+ "n": 69,
1690
+ "num_samples": 200
1691
+ },
1692
+ "n": 69
1693
+ },
1694
+ "proposal_oracle_utility_gain_over_base_at_8": {
1695
+ "macro_by_seed": {
1696
+ "ci": {
1697
+ "confidence": 0.95,
1698
+ "high": -0.025295809314236143,
1699
+ "low": -0.4126824418096043,
1700
+ "mean": -0.2097537833114043,
1701
+ "n": 47,
1702
+ "num_samples": 200
1703
+ },
1704
+ "mean": -0.2097537833114043,
1705
+ "per_seed": {
1706
+ "1025740212": -0.005973592400550842,
1707
+ "1055157354": 0.029977411031723022,
1708
+ "1055567618": -1.1599239706993103,
1709
+ "1113268940": 0.0009514838457107544,
1710
+ "1113371261": 0.0,
1711
+ "1176943004": -1.9034668058156967,
1712
+ "1189392917": 0.0,
1713
+ "1231780997": -0.0010674148797988892,
1714
+ "1271718007": 0.020097307860851288,
1715
+ "1273247072": -0.017969928681850433,
1716
+ "1358421325": -1.1598625779151917,
1717
+ "1375492786": -0.17808932065963745,
1718
+ "141984615": -0.01871856302022934,
1719
+ "1475440973": 0.0052315592765808105,
1720
+ "1483645336": 0.0,
1721
+ "1522272686": 0.010118722915649414,
1722
+ "1528740417": -0.043816983699798584,
1723
+ "1549324805": -0.05327957744399706,
1724
+ "1607773860": 0.009314656257629395,
1725
+ "1665739511": 0.0,
1726
+ "1747178030": -1.9129380583763123,
1727
+ "1881223807": -0.008863061666488647,
1728
+ "1919198188": -0.00294572114944458,
1729
+ "1991414747": -1.9150912910699844,
1730
+ "1997090716": -1.8682253062725067,
1731
+ "2015956957": -0.3538366474211216,
1732
+ "2054899126": 0.0,
1733
+ "211245842": 0.0440671443939209,
1734
+ "2141802816": -0.04327954351902008,
1735
+ "256770874": 0.037312328815460205,
1736
+ "278008524": -0.026284693740308285,
1737
+ "416183106": 0.30440694093704224,
1738
+ "534717141": -0.04306729882955551,
1739
+ "551504635": 1.2727532982826233,
1740
+ "565369392": -0.6359496836860975,
1741
+ "565621826": 0.0,
1742
+ "585904130": -0.04499624793728193,
1743
+ "626486813": -0.45377432741224766,
1744
+ "636680137": -0.021984238177537918,
1745
+ "706479600": 0.01815354824066162,
1746
+ "750448296": 0.0011509992182254791,
1747
+ "75070641": 0.022626057267189026,
1748
+ "767747424": -1.3600103557109833e-05,
1749
+ "77147264": 0.007970944046974182,
1750
+ "794181626": 0.0,
1751
+ "957101489": 0.3378151208162308,
1752
+ "998761162": -0.10695688426494598
1753
+ }
1754
+ },
1755
+ "macro_by_task": {
1756
+ "ci": {
1757
+ "confidence": 0.95,
1758
+ "high": 0.056164648957258353,
1759
+ "low": -0.29419493908245187,
1760
+ "mean": -0.11944086916525573,
1761
+ "n": 5,
1762
+ "num_samples": 200
1763
+ },
1764
+ "mean": -0.11944086916525573,
1765
+ "per_task": {
1766
+ "LiftPegUpright-v1": 0.20467972507079443,
1767
+ "PickCube-v1": -0.017969928681850433,
1768
+ "PullCube-v1": -0.2363623414079056,
1769
+ "PushCube-v1": -0.45069821178913116,
1770
+ "StackCube-v1": -0.09685358901818593
1771
+ }
1772
+ },
1773
+ "metric": "proposal_oracle_utility_gain_over_base_at_8",
1774
+ "micro": {
1775
+ "confidence": 0.95,
1776
+ "high": -0.08154426396325015,
1777
+ "low": -0.33253368065840955,
1778
+ "mean": -0.20489988079213578,
1779
+ "n": 69,
1780
+ "num_samples": 200
1781
+ },
1782
+ "n": 69
1783
+ },
1784
+ "selected_safety_label_known_at_8": {
1785
+ "macro_by_seed": {
1786
+ "ci": {
1787
+ "confidence": 0.95,
1788
+ "high": 1.0,
1789
+ "low": 1.0,
1790
+ "mean": 1.0,
1791
+ "n": 47,
1792
+ "num_samples": 200
1793
+ },
1794
+ "mean": 1.0,
1795
+ "per_seed": {
1796
+ "1025740212": 1.0,
1797
+ "1055157354": 1.0,
1798
+ "1055567618": 1.0,
1799
+ "1113268940": 1.0,
1800
+ "1113371261": 1.0,
1801
+ "1176943004": 1.0,
1802
+ "1189392917": 1.0,
1803
+ "1231780997": 1.0,
1804
+ "1271718007": 1.0,
1805
+ "1273247072": 1.0,
1806
+ "1358421325": 1.0,
1807
+ "1375492786": 1.0,
1808
+ "141984615": 1.0,
1809
+ "1475440973": 1.0,
1810
+ "1483645336": 1.0,
1811
+ "1522272686": 1.0,
1812
+ "1528740417": 1.0,
1813
+ "1549324805": 1.0,
1814
+ "1607773860": 1.0,
1815
+ "1665739511": 1.0,
1816
+ "1747178030": 1.0,
1817
+ "1881223807": 1.0,
1818
+ "1919198188": 1.0,
1819
+ "1991414747": 1.0,
1820
+ "1997090716": 1.0,
1821
+ "2015956957": 1.0,
1822
+ "2054899126": 1.0,
1823
+ "211245842": 1.0,
1824
+ "2141802816": 1.0,
1825
+ "256770874": 1.0,
1826
+ "278008524": 1.0,
1827
+ "416183106": 1.0,
1828
+ "534717141": 1.0,
1829
+ "551504635": 1.0,
1830
+ "565369392": 1.0,
1831
+ "565621826": 1.0,
1832
+ "585904130": 1.0,
1833
+ "626486813": 1.0,
1834
+ "636680137": 1.0,
1835
+ "706479600": 1.0,
1836
+ "750448296": 1.0,
1837
+ "75070641": 1.0,
1838
+ "767747424": 1.0,
1839
+ "77147264": 1.0,
1840
+ "794181626": 1.0,
1841
+ "957101489": 1.0,
1842
+ "998761162": 1.0
1843
+ }
1844
+ },
1845
+ "macro_by_task": {
1846
+ "ci": {
1847
+ "confidence": 0.95,
1848
+ "high": 1.0,
1849
+ "low": 1.0,
1850
+ "mean": 1.0,
1851
+ "n": 5,
1852
+ "num_samples": 200
1853
+ },
1854
+ "mean": 1.0,
1855
+ "per_task": {
1856
+ "LiftPegUpright-v1": 1.0,
1857
+ "PickCube-v1": 1.0,
1858
+ "PullCube-v1": 1.0,
1859
+ "PushCube-v1": 1.0,
1860
+ "StackCube-v1": 1.0
1861
+ }
1862
+ },
1863
+ "metric": "selected_safety_label_known_at_8",
1864
+ "micro": {
1865
+ "confidence": 0.95,
1866
+ "high": 1.0,
1867
+ "low": 1.0,
1868
+ "mean": 1.0,
1869
+ "n": 69,
1870
+ "num_samples": 200
1871
+ },
1872
+ "n": 69
1873
+ },
1874
+ "selected_success_at_8": {
1875
+ "macro_by_seed": {
1876
+ "ci": {
1877
+ "confidence": 0.95,
1878
+ "high": 0.2553191489361702,
1879
+ "low": 0.06382978723404255,
1880
+ "mean": 0.14893617021276595,
1881
+ "n": 47,
1882
+ "num_samples": 200
1883
+ },
1884
+ "mean": 0.14893617021276595,
1885
+ "per_seed": {
1886
+ "1025740212": 0.0,
1887
+ "1055157354": 0.0,
1888
+ "1055567618": 0.0,
1889
+ "1113268940": 0.0,
1890
+ "1113371261": 1.0,
1891
+ "1176943004": 0.0,
1892
+ "1189392917": 1.0,
1893
+ "1231780997": 0.0,
1894
+ "1271718007": 0.5,
1895
+ "1273247072": 0.0,
1896
+ "1358421325": 0.0,
1897
+ "1375492786": 0.0,
1898
+ "141984615": 0.0,
1899
+ "1475440973": 0.0,
1900
+ "1483645336": 0.0,
1901
+ "1522272686": 0.0,
1902
+ "1528740417": 0.0,
1903
+ "1549324805": 0.0,
1904
+ "1607773860": 0.0,
1905
+ "1665739511": 1.0,
1906
+ "1747178030": 0.0,
1907
+ "1881223807": 0.0,
1908
+ "1919198188": 0.0,
1909
+ "1991414747": 0.0,
1910
+ "1997090716": 0.0,
1911
+ "2015956957": 0.0,
1912
+ "2054899126": 1.0,
1913
+ "211245842": 0.0,
1914
+ "2141802816": 0.0,
1915
+ "256770874": 0.0,
1916
+ "278008524": 0.0,
1917
+ "416183106": 0.0,
1918
+ "534717141": 0.0,
1919
+ "551504635": 0.0,
1920
+ "565369392": 0.0,
1921
+ "565621826": 1.0,
1922
+ "585904130": 0.0,
1923
+ "626486813": 0.0,
1924
+ "636680137": 0.0,
1925
+ "706479600": 0.0,
1926
+ "750448296": 0.0,
1927
+ "75070641": 0.0,
1928
+ "767747424": 0.0,
1929
+ "77147264": 0.5,
1930
+ "794181626": 1.0,
1931
+ "957101489": 0.0,
1932
+ "998761162": 0.0
1933
+ }
1934
+ },
1935
+ "macro_by_task": {
1936
+ "ci": {
1937
+ "confidence": 0.95,
1938
+ "high": 0.24253393665158374,
1939
+ "low": 0.013333333333333332,
1940
+ "mean": 0.10202111613876322,
1941
+ "n": 5,
1942
+ "num_samples": 200
1943
+ },
1944
+ "mean": 0.10202111613876322,
1945
+ "per_task": {
1946
+ "LiftPegUpright-v1": 0.0,
1947
+ "PickCube-v1": 0.0,
1948
+ "PullCube-v1": 0.058823529411764705,
1949
+ "PushCube-v1": 0.38461538461538464,
1950
+ "StackCube-v1": 0.06666666666666667
1951
+ }
1952
+ },
1953
+ "metric": "selected_success_at_8",
1954
+ "micro": {
1955
+ "confidence": 0.95,
1956
+ "high": 0.21739130434782608,
1957
+ "low": 0.057971014492753624,
1958
+ "mean": 0.11594202898550725,
1959
+ "n": 69,
1960
+ "num_samples": 200
1961
+ },
1962
+ "n": 69
1963
+ },
1964
+ "selected_success_gain_over_base_at_8": {
1965
+ "macro_by_seed": {
1966
+ "ci": {
1967
+ "confidence": 0.95,
1968
+ "high": -0.0794326241134752,
1969
+ "low": -0.2822695035460993,
1970
+ "mean": -0.1726950354609929,
1971
+ "n": 47,
1972
+ "num_samples": 200
1973
+ },
1974
+ "mean": -0.1726950354609929,
1975
+ "per_seed": {
1976
+ "1025740212": 0.0,
1977
+ "1055157354": 0.0,
1978
+ "1055567618": -1.0,
1979
+ "1113268940": 0.0,
1980
+ "1113371261": 0.0,
1981
+ "1176943004": -1.0,
1982
+ "1189392917": 0.0,
1983
+ "1231780997": 0.0,
1984
+ "1271718007": 0.0,
1985
+ "1273247072": 0.0,
1986
+ "1358421325": -1.0,
1987
+ "1375492786": 0.0,
1988
+ "141984615": 0.0,
1989
+ "1475440973": 0.0,
1990
+ "1483645336": -1.0,
1991
+ "1522272686": 0.0,
1992
+ "1528740417": 0.0,
1993
+ "1549324805": -0.3333333333333333,
1994
+ "1607773860": 0.0,
1995
+ "1665739511": 0.0,
1996
+ "1747178030": -1.0,
1997
+ "1881223807": 0.0,
1998
+ "1919198188": 0.0,
1999
+ "1991414747": -1.0,
2000
+ "1997090716": -1.0,
2001
+ "2015956957": -0.2,
2002
+ "2054899126": 0.0,
2003
+ "211245842": 0.0,
2004
+ "2141802816": 0.0,
2005
+ "256770874": 0.0,
2006
+ "278008524": 0.0,
2007
+ "416183106": 0.0,
2008
+ "534717141": 0.0,
2009
+ "551504635": 0.0,
2010
+ "565369392": -0.3333333333333333,
2011
+ "565621826": 0.0,
2012
+ "585904130": 0.0,
2013
+ "626486813": -0.25,
2014
+ "636680137": 0.0,
2015
+ "706479600": 0.0,
2016
+ "750448296": 0.0,
2017
+ "75070641": 0.0,
2018
+ "767747424": 0.0,
2019
+ "77147264": 0.0,
2020
+ "794181626": 0.0,
2021
+ "957101489": 0.0,
2022
+ "998761162": 0.0
2023
+ }
2024
+ },
2025
+ "macro_by_task": {
2026
+ "ci": {
2027
+ "confidence": 0.95,
2028
+ "high": -0.029411764705882353,
2029
+ "low": -0.21402714932126696,
2030
+ "mean": -0.11761689291101056,
2031
+ "n": 5,
2032
+ "num_samples": 200
2033
+ },
2034
+ "mean": -0.11761689291101056,
2035
+ "per_task": {
2036
+ "LiftPegUpright-v1": 0.0,
2037
+ "PickCube-v1": 0.0,
2038
+ "PullCube-v1": -0.14705882352941177,
2039
+ "PushCube-v1": -0.3076923076923077,
2040
+ "StackCube-v1": -0.13333333333333333
2041
+ }
2042
+ },
2043
+ "metric": "selected_success_gain_over_base_at_8",
2044
+ "micro": {
2045
+ "confidence": 0.95,
2046
+ "high": -0.08695652173913043,
2047
+ "low": -0.2463768115942029,
2048
+ "mean": -0.15942028985507245,
2049
+ "n": 69,
2050
+ "num_samples": 200
2051
+ },
2052
+ "n": 69
2053
+ },
2054
+ "selected_unsafe_known_at_8": {
2055
+ "macro_by_seed": {
2056
+ "ci": {
2057
+ "confidence": 0.95,
2058
+ "high": 0.2315602836879433,
2059
+ "low": 0.07943262411347518,
2060
+ "mean": 0.15567375886524823,
2061
+ "n": 47,
2062
+ "num_samples": 200
2063
+ },
2064
+ "mean": 0.15567375886524823,
2065
+ "per_seed": {
2066
+ "1025740212": 0.0,
2067
+ "1055157354": 0.0,
2068
+ "1055567618": 0.0,
2069
+ "1113268940": 0.0,
2070
+ "1113371261": 0.0,
2071
+ "1176943004": 0.0,
2072
+ "1189392917": 0.0,
2073
+ "1231780997": 0.0,
2074
+ "1271718007": 0.0,
2075
+ "1273247072": 0.0,
2076
+ "1358421325": 0.0,
2077
+ "1375492786": 1.0,
2078
+ "141984615": 0.3333333333333333,
2079
+ "1475440973": 0.0,
2080
+ "1483645336": 0.0,
2081
+ "1522272686": 0.0,
2082
+ "1528740417": 0.0,
2083
+ "1549324805": 0.3333333333333333,
2084
+ "1607773860": 0.0,
2085
+ "1665739511": 0.0,
2086
+ "1747178030": 1.0,
2087
+ "1881223807": 0.0,
2088
+ "1919198188": 0.0,
2089
+ "1991414747": 0.0,
2090
+ "1997090716": 0.0,
2091
+ "2015956957": 0.4,
2092
+ "2054899126": 1.0,
2093
+ "211245842": 0.0,
2094
+ "2141802816": 0.0,
2095
+ "256770874": 0.0,
2096
+ "278008524": 0.5,
2097
+ "416183106": 0.0,
2098
+ "534717141": 0.0,
2099
+ "551504635": 0.0,
2100
+ "565369392": 0.6666666666666666,
2101
+ "565621826": 1.0,
2102
+ "585904130": 0.3333333333333333,
2103
+ "626486813": 0.25,
2104
+ "636680137": 0.0,
2105
+ "706479600": 0.0,
2106
+ "750448296": 0.0,
2107
+ "75070641": 0.0,
2108
+ "767747424": 0.5,
2109
+ "77147264": 0.0,
2110
+ "794181626": 0.0,
2111
+ "957101489": 0.0,
2112
+ "998761162": 0.0
2113
+ }
2114
+ },
2115
+ "macro_by_task": {
2116
+ "ci": {
2117
+ "confidence": 0.95,
2118
+ "high": 0.2647058823529412,
2119
+ "low": 0.0,
2120
+ "mean": 0.08823529411764705,
2121
+ "n": 5,
2122
+ "num_samples": 200
2123
+ },
2124
+ "mean": 0.08823529411764705,
2125
+ "per_task": {
2126
+ "LiftPegUpright-v1": 0.0,
2127
+ "PickCube-v1": 0.0,
2128
+ "PullCube-v1": 0.4411764705882353,
2129
+ "PushCube-v1": 0.0,
2130
+ "StackCube-v1": 0.0
2131
+ }
2132
+ },
2133
+ "metric": "selected_unsafe_known_at_8",
2134
+ "micro": {
2135
+ "confidence": 0.95,
2136
+ "high": 0.30434782608695654,
2137
+ "low": 0.14492753623188406,
2138
+ "mean": 0.21739130434782608,
2139
+ "n": 69,
2140
+ "num_samples": 200
2141
+ },
2142
+ "n": 69
2143
+ },
2144
+ "selected_utility_at_8": {
2145
+ "macro_by_seed": {
2146
+ "ci": {
2147
+ "confidence": 0.95,
2148
+ "high": 0.681915133363891,
2149
+ "low": 0.32914034059793024,
2150
+ "mean": 0.49882138063800885,
2151
+ "n": 47,
2152
+ "num_samples": 200
2153
+ },
2154
+ "mean": 0.49882138063800885,
2155
+ "per_seed": {
2156
+ "1025740212": 0.2067173719406128,
2157
+ "1055157354": 0.20325151085853577,
2158
+ "1055567618": 0.8212382197380066,
2159
+ "1113268940": 0.2133198231458664,
2160
+ "1113371261": 2.0,
2161
+ "1176943004": 0.06349706649780273,
2162
+ "1189392917": 2.0,
2163
+ "1231780997": 0.2103814333677292,
2164
+ "1271718007": 1.1248890534043312,
2165
+ "1273247072": 0.08980755507946014,
2166
+ "1358421325": 0.15342360734939575,
2167
+ "1375492786": 0.003411312820389867,
2168
+ "141984615": 0.08821865008212626,
2169
+ "1475440973": 0.7232699394226074,
2170
+ "1483645336": 0.15040680766105652,
2171
+ "1522272686": 0.2244599312543869,
2172
+ "1528740417": 0.26118800044059753,
2173
+ "1549324805": 0.0908410723010699,
2174
+ "1607773860": 0.8011545538902283,
2175
+ "1665739511": 2.0,
2176
+ "1747178030": 0.0769355297088623,
2177
+ "1881223807": 0.04914247989654541,
2178
+ "1919198188": 0.05640652775764465,
2179
+ "1991414747": 0.08029218018054962,
2180
+ "1997090716": 0.12443000078201294,
2181
+ "2015956957": 0.10171020515263081,
2182
+ "2054899126": 2.0,
2183
+ "211245842": 0.7903563976287842,
2184
+ "2141802816": 0.15455174446105957,
2185
+ "256770874": 0.17646989226341248,
2186
+ "278008524": 0.026269666850566864,
2187
+ "416183106": 0.5344269275665283,
2188
+ "534717141": 0.08970203250646591,
2189
+ "551504635": 0.7269402146339417,
2190
+ "565369392": 0.20857956757148108,
2191
+ "565621826": 2.0,
2192
+ "585904130": 0.06864311297734578,
2193
+ "626486813": 0.08136882539838552,
2194
+ "636680137": 0.026473185047507286,
2195
+ "706479600": 0.8187581300735474,
2196
+ "750448296": 0.1222188975661993,
2197
+ "75070641": 0.192473903298378,
2198
+ "767747424": 0.09243520395830274,
2199
+ "77147264": 1.0893699154257774,
2200
+ "794181626": 2.0,
2201
+ "957101489": 0.23480956256389618,
2202
+ "998761162": 0.09236487746238708
2203
+ }
2204
+ },
2205
+ "macro_by_task": {
2206
+ "ci": {
2207
+ "confidence": 0.95,
2208
+ "high": 0.6366648905951849,
2209
+ "low": 0.19669590667620593,
2210
+ "mean": 0.40191000551159517,
2211
+ "n": 5,
2212
+ "num_samples": 200
2213
+ },
2214
+ "mean": 0.40191000551159517,
2215
+ "per_task": {
2216
+ "LiftPegUpright-v1": 0.39030279715855914,
2217
+ "PickCube-v1": 0.08980755507946014,
2218
+ "PullCube-v1": 0.20678081303177512,
2219
+ "PushCube-v1": 0.8269423888279841,
2220
+ "StackCube-v1": 0.4957164734601974
2221
+ }
2222
+ },
2223
+ "metric": "selected_utility_at_8",
2224
+ "micro": {
2225
+ "confidence": 0.95,
2226
+ "high": 0.5762512336005929,
2227
+ "low": 0.262371721232067,
2228
+ "mean": 0.4006981179388105,
2229
+ "n": 69,
2230
+ "num_samples": 200
2231
+ },
2232
+ "n": 69
2233
+ },
2234
+ "selected_utility_gain_over_base_at_8": {
2235
+ "macro_by_seed": {
2236
+ "ci": {
2237
+ "confidence": 0.95,
2238
+ "high": -0.15531579189544134,
2239
+ "low": -0.4976459303246613,
2240
+ "mean": -0.32850989265142816,
2241
+ "n": 47,
2242
+ "num_samples": 200
2243
+ },
2244
+ "mean": -0.32850989265142816,
2245
+ "per_seed": {
2246
+ "1025740212": -0.006895139813423157,
2247
+ "1055157354": 0.015253812074661255,
2248
+ "1055567618": -1.1787617802619934,
2249
+ "1113268940": -0.00569605827331543,
2250
+ "1113371261": 0.0,
2251
+ "1176943004": -1.9365029335021973,
2252
+ "1189392917": 0.0,
2253
+ "1231780997": -0.0010674148797988892,
2254
+ "1271718007": 0.020097307860851288,
2255
+ "1273247072": -0.029644250869750977,
2256
+ "1358421325": -1.8465763926506042,
2257
+ "1375492786": -0.17982337879948318,
2258
+ "141984615": -0.03661966983539363,
2259
+ "1475440973": 0.0002951622009277344,
2260
+ "1483645336": -1.8495931923389435,
2261
+ "1522272686": -0.3285377770662308,
2262
+ "1528740417": -0.06528261303901672,
2263
+ "1549324805": -0.6723979264497757,
2264
+ "1607773860": -0.0021181106567382812,
2265
+ "1665739511": 0.0,
2266
+ "1747178030": -1.9230644702911377,
2267
+ "1881223807": -0.010729014873504639,
2268
+ "1919198188": -0.009721159934997559,
2269
+ "1991414747": -1.9197078198194504,
2270
+ "1997090716": -1.875569999217987,
2271
+ "2015956957": -0.3712019260972738,
2272
+ "2054899126": 0.0,
2273
+ "211245842": -0.0009227395057678223,
2274
+ "2141802816": -0.0820978432893753,
2275
+ "256770874": -0.00513109564781189,
2276
+ "278008524": -0.03438602015376091,
2277
+ "416183106": 0.30440694093704224,
2278
+ "534717141": -0.04306729882955551,
2279
+ "551504635": -0.0003064870834350586,
2280
+ "565369392": -0.6359496836860975,
2281
+ "565621826": 0.0,
2282
+ "585904130": -0.06777803599834442,
2283
+ "626486813": -0.4638675441965461,
2284
+ "636680137": -0.0437681470066309,
2285
+ "706479600": -0.00013238191604614258,
2286
+ "750448296": -0.024850530549883842,
2287
+ "75070641": -0.001552477478981018,
2288
+ "767747424": -0.016510562505573034,
2289
+ "77147264": -0.0009331181645393372,
2290
+ "794181626": 0.0,
2291
+ "957101489": 0.0026340484619140625,
2292
+ "998761162": -0.11188723146915436
2293
+ }
2294
+ },
2295
+ "macro_by_task": {
2296
+ "ci": {
2297
+ "confidence": 0.95,
2298
+ "high": -0.07772063214477955,
2299
+ "low": -0.42360685428416256,
2300
+ "mean": -0.22945633145911706,
2301
+ "n": 5,
2302
+ "num_samples": 200
2303
+ },
2304
+ "mean": -0.22945633145911706,
2305
+ "per_task": {
2306
+ "LiftPegUpright-v1": -0.013158758481343588,
2307
+ "PickCube-v1": -0.029644250869750977,
2308
+ "PullCube-v1": -0.30299714202170863,
2309
+ "PushCube-v1": -0.6006261236392535,
2310
+ "StackCube-v1": -0.20085538228352864
2311
+ }
2312
+ },
2313
+ "metric": "selected_utility_gain_over_base_at_8",
2314
+ "micro": {
2315
+ "confidence": 0.95,
2316
+ "high": -0.16835133358836174,
2317
+ "low": -0.47174493824401736,
2318
+ "mean": -0.3077024633631758,
2319
+ "n": 69,
2320
+ "num_samples": 200
2321
+ },
2322
+ "n": 69
2323
+ },
2324
+ "selector_gap_fraction_to_hidden_at_8": {
2325
+ "macro_by_seed": {
2326
+ "ci": {
2327
+ "confidence": 0.95,
2328
+ "high": 0.5919580203395092,
2329
+ "low": 0.09161768413805887,
2330
+ "mean": 0.3003313627738701,
2331
+ "n": 34,
2332
+ "num_samples": 200
2333
+ },
2334
+ "mean": 0.3003313627738701,
2335
+ "per_seed": {
2336
+ "1025740212": 0.000515872070833396,
2337
+ "1055157354": 0.00812559616904461,
2338
+ "1113268940": 0.0037325106101020683,
2339
+ "1231780997": 0.0,
2340
+ "1271718007": 0.0,
2341
+ "1273247072": 0.006207935656652067,
2342
+ "1375492786": 0.028797165851366206,
2343
+ "141984615": 0.6932408439097882,
2344
+ "1475440973": 0.0038655439122129136,
2345
+ "1522272686": 0.23404005780034906,
2346
+ "1528740417": 0.01282656254029075,
2347
+ "1549324805": 0.30888217864267864,
2348
+ "1607773860": 0.009553359880460726,
2349
+ "1881223807": 0.0009617678426897979,
2350
+ "1919198188": 0.003503560572450107,
2351
+ "2015956957": 0.24547848203994205,
2352
+ "211245842": 0.0372210700434455,
2353
+ "2141802816": 0.022013945441980202,
2354
+ "256770874": 0.023341095205786425,
2355
+ "278008524": 0.024682996062310017,
2356
+ "416183106": 0.0,
2357
+ "534717141": 0.0,
2358
+ "551504635": 1.0002408063556767,
2359
+ "565369392": 0.0,
2360
+ "585904130": 0.37677313988714195,
2361
+ "626486813": 0.3592836940693531,
2362
+ "636680137": 0.011288410924605892,
2363
+ "706479600": 0.015481994126988717,
2364
+ "750448296": 3.5662633342346166,
2365
+ "75070641": 0.01338808855722787,
2366
+ "767747424": 2.999421934457336,
2367
+ "77147264": 0.009787943172609816,
2368
+ "957101489": 0.18960087668660824,
2369
+ "998761162": 0.002745567587037426
2370
+ }
2371
+ },
2372
+ "macro_by_task": {
2373
+ "ci": {
2374
+ "confidence": 0.95,
2375
+ "high": 0.5931182317297116,
2376
+ "low": 0.02185954946782361,
2377
+ "mean": 0.23180818928297858,
2378
+ "n": 5,
2379
+ "num_samples": 200
2380
+ },
2381
+ "mean": 0.23180818928297858,
2382
+ "per_task": {
2383
+ "LiftPegUpright-v1": 0.1701968825815193,
2384
+ "PickCube-v1": 0.006207935656652067,
2385
+ "PullCube-v1": 0.9297287801367955,
2386
+ "PushCube-v1": 0.007570377855345064,
2387
+ "StackCube-v1": 0.04533697018458092
2388
+ }
2389
+ },
2390
+ "metric": "selector_gap_fraction_to_hidden_at_8",
2391
+ "micro": {
2392
+ "confidence": 0.95,
2393
+ "high": 1.1723803267231159,
2394
+ "low": 0.13113286877419325,
2395
+ "mean": 0.5340878289695119,
2396
+ "n": 50,
2397
+ "num_samples": 200
2398
+ },
2399
+ "n": 50
2400
+ },
2401
+ "selector_regret_at_8": {
2402
+ "macro_by_seed": {
2403
+ "ci": {
2404
+ "confidence": 0.95,
2405
+ "high": 0.2221361303896848,
2406
+ "low": 0.03224338199580022,
2407
+ "mean": 0.11875610934002392,
2408
+ "n": 47,
2409
+ "num_samples": 200
2410
+ },
2411
+ "mean": 0.11875610934002392,
2412
+ "per_seed": {
2413
+ "1025740212": 0.0009215474128723145,
2414
+ "1055157354": 0.014723598957061768,
2415
+ "1055567618": 0.018837809562683105,
2416
+ "1113268940": 0.006647542119026184,
2417
+ "1113371261": 0.0,
2418
+ "1176943004": 0.03303612768650055,
2419
+ "1189392917": 0.0,
2420
+ "1231780997": 0.0,
2421
+ "1271718007": 0.0,
2422
+ "1273247072": 0.011674322187900543,
2423
+ "1358421325": 0.6867138147354126,
2424
+ "1375492786": 0.0017340581398457289,
2425
+ "141984615": 0.017901106815164287,
2426
+ "1475440973": 0.004936397075653076,
2427
+ "1483645336": 1.8495931923389435,
2428
+ "1522272686": 0.3386564999818802,
2429
+ "1528740417": 0.02146562933921814,
2430
+ "1549324805": 0.6191183490057787,
2431
+ "1607773860": 0.011432766914367676,
2432
+ "1665739511": 0.0,
2433
+ "1747178030": 0.01012641191482544,
2434
+ "1881223807": 0.0018659532070159912,
2435
+ "1919198188": 0.0067754387855529785,
2436
+ "1991414747": 0.004616528749465942,
2437
+ "1997090716": 0.007344692945480347,
2438
+ "2015956957": 0.01736527867615223,
2439
+ "2054899126": 0.0,
2440
+ "211245842": 0.04498988389968872,
2441
+ "2141802816": 0.038818299770355225,
2442
+ "256770874": 0.042443424463272095,
2443
+ "278008524": 0.008101326413452625,
2444
+ "416183106": 0.0,
2445
+ "534717141": 0.0,
2446
+ "551504635": 1.2730597853660583,
2447
+ "565369392": 0.0,
2448
+ "565621826": 0.0,
2449
+ "585904130": 0.022781788061062496,
2450
+ "626486813": 0.01009321678429842,
2451
+ "636680137": 0.02178390882909298,
2452
+ "706479600": 0.018285930156707764,
2453
+ "750448296": 0.02600152976810932,
2454
+ "75070641": 0.024178534746170044,
2455
+ "767747424": 0.016496962402015924,
2456
+ "77147264": 0.00890406221151352,
2457
+ "794181626": 0.0,
2458
+ "957101489": 0.3351810723543167,
2459
+ "998761162": 0.004930347204208374
2460
+ }
2461
+ },
2462
+ "macro_by_task": {
2463
+ "ci": {
2464
+ "confidence": 0.95,
2465
+ "high": 0.16483040890711195,
2466
+ "low": 0.04860531061887741,
2467
+ "mean": 0.11001546229386132,
2468
+ "n": 5,
2469
+ "num_samples": 200
2470
+ },
2471
+ "mean": 0.11001546229386132,
2472
+ "per_task": {
2473
+ "LiftPegUpright-v1": 0.217838483552138,
2474
+ "PickCube-v1": 0.011674322187900543,
2475
+ "PullCube-v1": 0.06663480061380302,
2476
+ "PushCube-v1": 0.1499279118501223,
2477
+ "StackCube-v1": 0.10400179326534272
2478
+ }
2479
+ },
2480
+ "metric": "selector_regret_at_8",
2481
+ "micro": {
2482
+ "confidence": 0.95,
2483
+ "high": 0.1787753005552551,
2484
+ "low": 0.03932376045976644,
2485
+ "mean": 0.10280258257104002,
2486
+ "n": 69,
2487
+ "num_samples": 200
2488
+ },
2489
+ "n": 69
2490
+ },
2491
+ "success_selector_gap_at_8": {
2492
+ "macro_by_seed": {
2493
+ "ci": {
2494
+ "confidence": 0.95,
2495
+ "high": 0.11347517730496454,
2496
+ "low": 0.0,
2497
+ "mean": 0.04964539007092198,
2498
+ "n": 47,
2499
+ "num_samples": 200
2500
+ },
2501
+ "mean": 0.04964539007092198,
2502
+ "per_seed": {
2503
+ "1025740212": 0.0,
2504
+ "1055157354": 0.0,
2505
+ "1055567618": 0.0,
2506
+ "1113268940": 0.0,
2507
+ "1113371261": 0.0,
2508
+ "1176943004": 0.0,
2509
+ "1189392917": 0.0,
2510
+ "1231780997": 0.0,
2511
+ "1271718007": 0.0,
2512
+ "1273247072": 0.0,
2513
+ "1358421325": 0.0,
2514
+ "1375492786": 0.0,
2515
+ "141984615": 0.0,
2516
+ "1475440973": 0.0,
2517
+ "1483645336": 1.0,
2518
+ "1522272686": 0.0,
2519
+ "1528740417": 0.0,
2520
+ "1549324805": 0.3333333333333333,
2521
+ "1607773860": 0.0,
2522
+ "1665739511": 0.0,
2523
+ "1747178030": 0.0,
2524
+ "1881223807": 0.0,
2525
+ "1919198188": 0.0,
2526
+ "1991414747": 0.0,
2527
+ "1997090716": 0.0,
2528
+ "2015956957": 0.0,
2529
+ "2054899126": 0.0,
2530
+ "211245842": 0.0,
2531
+ "2141802816": 0.0,
2532
+ "256770874": 0.0,
2533
+ "278008524": 0.0,
2534
+ "416183106": 0.0,
2535
+ "534717141": 0.0,
2536
+ "551504635": 1.0,
2537
+ "565369392": 0.0,
2538
+ "565621826": 0.0,
2539
+ "585904130": 0.0,
2540
+ "626486813": 0.0,
2541
+ "636680137": 0.0,
2542
+ "706479600": 0.0,
2543
+ "750448296": 0.0,
2544
+ "75070641": 0.0,
2545
+ "767747424": 0.0,
2546
+ "77147264": 0.0,
2547
+ "794181626": 0.0,
2548
+ "957101489": 0.0,
2549
+ "998761162": 0.0
2550
+ }
2551
+ },
2552
+ "macro_by_task": {
2553
+ "ci": {
2554
+ "confidence": 0.95,
2555
+ "high": 0.11282051282051282,
2556
+ "low": 0.0058823529411764705,
2557
+ "mean": 0.054600301659125196,
2558
+ "n": 5,
2559
+ "num_samples": 200
2560
+ },
2561
+ "mean": 0.054600301659125196,
2562
+ "per_task": {
2563
+ "LiftPegUpright-v1": 0.16666666666666666,
2564
+ "PickCube-v1": 0.0,
2565
+ "PullCube-v1": 0.029411764705882353,
2566
+ "PushCube-v1": 0.07692307692307693,
2567
+ "StackCube-v1": 0.0
2568
+ }
2569
+ },
2570
+ "metric": "success_selector_gap_at_8",
2571
+ "micro": {
2572
+ "confidence": 0.95,
2573
+ "high": 0.08695652173913043,
2574
+ "low": 0.014492753623188406,
2575
+ "mean": 0.043478260869565216,
2576
+ "n": 69,
2577
+ "num_samples": 200
2578
+ },
2579
+ "n": 69
2580
+ },
2581
+ "success_support_gap_at_8": {
2582
+ "macro_by_seed": {
2583
+ "ci": {
2584
+ "confidence": 0.95,
2585
+ "high": 0.7340425531914894,
2586
+ "low": 0.526595744680851,
2587
+ "mean": 0.6372340425531916,
2588
+ "n": 47,
2589
+ "num_samples": 200
2590
+ },
2591
+ "mean": 0.6372340425531916,
2592
+ "per_seed": {
2593
+ "1025740212": 1.0,
2594
+ "1055157354": 1.0,
2595
+ "1055567618": 1.0,
2596
+ "1113268940": 1.0,
2597
+ "1113371261": 0.0,
2598
+ "1176943004": 1.0,
2599
+ "1189392917": 0.0,
2600
+ "1231780997": 1.0,
2601
+ "1271718007": 0.5,
2602
+ "1273247072": 1.0,
2603
+ "1358421325": 1.0,
2604
+ "1375492786": 0.0,
2605
+ "141984615": 0.0,
2606
+ "1475440973": 1.0,
2607
+ "1483645336": 0.0,
2608
+ "1522272686": 1.0,
2609
+ "1528740417": 1.0,
2610
+ "1549324805": 0.3333333333333333,
2611
+ "1607773860": 1.0,
2612
+ "1665739511": 0.0,
2613
+ "1747178030": 1.0,
2614
+ "1881223807": 1.0,
2615
+ "1919198188": 1.0,
2616
+ "1991414747": 1.0,
2617
+ "1997090716": 1.0,
2618
+ "2015956957": 0.2,
2619
+ "2054899126": 0.0,
2620
+ "211245842": 1.0,
2621
+ "2141802816": 1.0,
2622
+ "256770874": 1.0,
2623
+ "278008524": 0.5,
2624
+ "416183106": 1.0,
2625
+ "534717141": 1.0,
2626
+ "551504635": 0.0,
2627
+ "565369392": 0.6666666666666666,
2628
+ "565621826": 0.0,
2629
+ "585904130": 0.0,
2630
+ "626486813": 0.25,
2631
+ "636680137": 1.0,
2632
+ "706479600": 1.0,
2633
+ "750448296": 0.0,
2634
+ "75070641": 1.0,
2635
+ "767747424": 0.0,
2636
+ "77147264": 0.5,
2637
+ "794181626": 0.0,
2638
+ "957101489": 1.0,
2639
+ "998761162": 1.0
2640
+ }
2641
+ },
2642
+ "macro_by_task": {
2643
+ "ci": {
2644
+ "confidence": 0.95,
2645
+ "high": 0.9266666666666667,
2646
+ "low": 0.448868778280543,
2647
+ "mean": 0.7080844645550528,
2648
+ "n": 5,
2649
+ "num_samples": 200
2650
+ },
2651
+ "mean": 0.7080844645550528,
2652
+ "per_task": {
2653
+ "LiftPegUpright-v1": 0.8333333333333334,
2654
+ "PickCube-v1": 1.0,
2655
+ "PullCube-v1": 0.23529411764705882,
2656
+ "PushCube-v1": 0.5384615384615384,
2657
+ "StackCube-v1": 0.9333333333333333
2658
+ }
2659
+ },
2660
+ "metric": "success_support_gap_at_8",
2661
+ "micro": {
2662
+ "confidence": 0.95,
2663
+ "high": 0.6231884057971014,
2664
+ "low": 0.37681159420289856,
2665
+ "mean": 0.5072463768115942,
2666
+ "n": 69,
2667
+ "num_samples": 200
2668
+ },
2669
+ "n": 69
2670
+ },
2671
+ "success_total_car_to_hidden_at_8": {
2672
+ "macro_by_seed": {
2673
+ "ci": {
2674
+ "confidence": 0.95,
2675
+ "high": 0.7890070921985815,
2676
+ "low": 0.575177304964539,
2677
+ "mean": 0.6868794326241134,
2678
+ "n": 47,
2679
+ "num_samples": 200
2680
+ },
2681
+ "mean": 0.6868794326241134,
2682
+ "per_seed": {
2683
+ "1025740212": 1.0,
2684
+ "1055157354": 1.0,
2685
+ "1055567618": 1.0,
2686
+ "1113268940": 1.0,
2687
+ "1113371261": 0.0,
2688
+ "1176943004": 1.0,
2689
+ "1189392917": 0.0,
2690
+ "1231780997": 1.0,
2691
+ "1271718007": 0.5,
2692
+ "1273247072": 1.0,
2693
+ "1358421325": 1.0,
2694
+ "1375492786": 0.0,
2695
+ "141984615": 0.0,
2696
+ "1475440973": 1.0,
2697
+ "1483645336": 1.0,
2698
+ "1522272686": 1.0,
2699
+ "1528740417": 1.0,
2700
+ "1549324805": 0.6666666666666666,
2701
+ "1607773860": 1.0,
2702
+ "1665739511": 0.0,
2703
+ "1747178030": 1.0,
2704
+ "1881223807": 1.0,
2705
+ "1919198188": 1.0,
2706
+ "1991414747": 1.0,
2707
+ "1997090716": 1.0,
2708
+ "2015956957": 0.2,
2709
+ "2054899126": 0.0,
2710
+ "211245842": 1.0,
2711
+ "2141802816": 1.0,
2712
+ "256770874": 1.0,
2713
+ "278008524": 0.5,
2714
+ "416183106": 1.0,
2715
+ "534717141": 1.0,
2716
+ "551504635": 1.0,
2717
+ "565369392": 0.6666666666666666,
2718
+ "565621826": 0.0,
2719
+ "585904130": 0.0,
2720
+ "626486813": 0.25,
2721
+ "636680137": 1.0,
2722
+ "706479600": 1.0,
2723
+ "750448296": 0.0,
2724
+ "75070641": 1.0,
2725
+ "767747424": 0.0,
2726
+ "77147264": 0.5,
2727
+ "794181626": 0.0,
2728
+ "957101489": 1.0,
2729
+ "998761162": 1.0
2730
+ }
2731
+ },
2732
+ "macro_by_task": {
2733
+ "ci": {
2734
+ "confidence": 0.95,
2735
+ "high": 0.9733333333333334,
2736
+ "low": 0.48190045248868774,
2737
+ "mean": 0.7626847662141779,
2738
+ "n": 5,
2739
+ "num_samples": 200
2740
+ },
2741
+ "mean": 0.7626847662141779,
2742
+ "per_task": {
2743
+ "LiftPegUpright-v1": 1.0,
2744
+ "PickCube-v1": 1.0,
2745
+ "PullCube-v1": 0.2647058823529412,
2746
+ "PushCube-v1": 0.6153846153846154,
2747
+ "StackCube-v1": 0.9333333333333333
2748
+ }
2749
+ },
2750
+ "metric": "success_total_car_to_hidden_at_8",
2751
+ "micro": {
2752
+ "confidence": 0.95,
2753
+ "high": 0.6811594202898551,
2754
+ "low": 0.42028985507246375,
2755
+ "mean": 0.5507246376811594,
2756
+ "n": 69,
2757
+ "num_samples": 200
2758
+ },
2759
+ "n": 69
2760
+ },
2761
+ "support_gap_at_8": {
2762
+ "macro_by_seed": {
2763
+ "ci": {
2764
+ "confidence": 0.95,
2765
+ "high": 1.2656197987483642,
2766
+ "low": 0.9043001213196848,
2767
+ "mean": 1.0893819265939453,
2768
+ "n": 47,
2769
+ "num_samples": 200
2770
+ },
2771
+ "mean": 1.0893819265939453,
2772
+ "per_seed": {
2773
+ "1025740212": 1.792361080646515,
2774
+ "1055157354": 1.7820248901844025,
2775
+ "1055567618": 1.1599239706993103,
2776
+ "1113268940": 1.7800326347351074,
2777
+ "1113371261": 0.0,
2778
+ "1176943004": 1.9034668058156967,
2779
+ "1189392917": 0.0,
2780
+ "1231780997": 1.7896185666322708,
2781
+ "1271718007": 0.8751109465956688,
2782
+ "1273247072": 1.8985181227326393,
2783
+ "1358421325": 1.1598625779151917,
2784
+ "1375492786": 0.23830559849739075,
2785
+ "141984615": 0.11338664591312408,
2786
+ "1475440973": 1.2717936635017395,
2787
+ "1483645336": 0.0,
2788
+ "1522272686": 1.436883568763733,
2789
+ "1528740417": 1.7173463702201843,
2790
+ "1549324805": 0.6927911005914211,
2791
+ "1607773860": 1.187412679195404,
2792
+ "1665739511": 0.0,
2793
+ "1747178030": 1.9129380583763123,
2794
+ "1881223807": 1.9489915668964386,
2795
+ "1919198188": 1.9368180334568024,
2796
+ "1991414747": 1.9150912910699844,
2797
+ "1997090716": 1.8682253062725067,
2798
+ "2015956957": 0.4130037985742092,
2799
+ "2054899126": 0.0,
2800
+ "211245842": 1.164653718471527,
2801
+ "2141802816": 1.8066299557685852,
2802
+ "256770874": 1.7810866832733154,
2803
+ "278008524": 1.0794402016326785,
2804
+ "416183106": 1.4655730724334717,
2805
+ "534717141": 1.910297967493534,
2806
+ "551504635": 0.0,
2807
+ "565369392": 1.218052012224992,
2808
+ "565621826": 0.0,
2809
+ "585904130": 0.14622209096948305,
2810
+ "626486813": 0.5353103466331959,
2811
+ "636680137": 1.9517429061233997,
2812
+ "706479600": 1.1629559397697449,
2813
+ "750448296": 0.08758716285228729,
2814
+ "75070641": 1.783347561955452,
2815
+ "767747424": 0.0797034902498126,
2816
+ "77147264": 0.901726022362709,
2817
+ "794181626": 0.0,
2818
+ "957101489": 1.430009365081787,
2819
+ "998761162": 1.9027047753334045
2820
+ }
2821
+ },
2822
+ "macro_by_task": {
2823
+ "ci": {
2824
+ "confidence": 0.95,
2825
+ "high": 1.6221085573618228,
2826
+ "low": 0.8659844333174475,
2827
+ "mean": 1.2452764358585575,
2828
+ "n": 5,
2829
+ "num_samples": 200
2830
+ },
2831
+ "mean": 1.2452764358585575,
2832
+ "per_task": {
2833
+ "LiftPegUpright-v1": 1.3918587192893028,
2834
+ "PickCube-v1": 1.8985181227326393,
2835
+ "PullCube-v1": 0.5125939046744915,
2836
+ "PushCube-v1": 1.0231296993218935,
2837
+ "StackCube-v1": 1.40028173327446
2838
+ }
2839
+ },
2840
+ "metric": "support_gap_at_8",
2841
+ "micro": {
2842
+ "confidence": 0.95,
2843
+ "high": 1.076073990891809,
2844
+ "low": 0.6788255989119627,
2845
+ "mean": 0.8983010911261258,
2846
+ "n": 69,
2847
+ "num_samples": 200
2848
+ },
2849
+ "n": 69
2850
+ },
2851
+ "support_gap_fraction_to_hidden_at_8": {
2852
+ "macro_by_seed": {
2853
+ "ci": {
2854
+ "confidence": 0.95,
2855
+ "high": 1.5908501015659637,
2856
+ "low": 1.0551446395118897,
2857
+ "mean": 1.3055196354245542,
2858
+ "n": 34,
2859
+ "num_samples": 200
2860
+ },
2861
+ "mean": 1.3055196354245542,
2862
+ "per_seed": {
2863
+ "1025740212": 1.0033439510967557,
2864
+ "1055157354": 0.9834561959377184,
2865
+ "1113268940": 0.9994657538852907,
2866
+ "1231780997": 1.0005968042226543,
2867
+ "1271718007": 0.9775501312004184,
2868
+ "1273247072": 1.0095556863358777,
2869
+ "1375492786": 3.957494668459602,
2870
+ "141984615": 2.951807933296765,
2871
+ "1475440973": 0.9959033234557427,
2872
+ "1522272686": 0.9930071134138249,
2873
+ "1528740417": 1.026182380813105,
2874
+ "1549324805": 2.1222703900724977,
2875
+ "1607773860": 0.9922165592932636,
2876
+ "1881223807": 1.0045682858854952,
2877
+ "1919198188": 1.00152322422256,
2878
+ "2015956957": 1.2234746297648738,
2879
+ "211245842": 0.9635423316104156,
2880
+ "2141802816": 1.0245439268442436,
2881
+ "256770874": 0.9794806703218564,
2882
+ "278008524": 1.1234960169117114,
2883
+ "416183106": 0.8280167354221304,
2884
+ "534717141": 1.0230647983413697,
2885
+ "551504635": 0.0,
2886
+ "565369392": 4.740673332898252,
2887
+ "585904130": 1.2148487255248963,
2888
+ "626486813": 0.7858811145047201,
2889
+ "636680137": 1.0113922214952087,
2890
+ "706479600": 0.9846300885523818,
2891
+ "750448296": 1.41259115683824,
2892
+ "75070641": 0.987471546081095,
2893
+ "767747424": 2.2059097841347444,
2894
+ "77147264": 0.9912378029813232,
2895
+ "957101489": 0.8089091289825484,
2896
+ "998761162": 1.0595611916332572
2897
+ }
2898
+ },
2899
+ "macro_by_task": {
2900
+ "ci": {
2901
+ "confidence": 0.95,
2902
+ "high": 1.559810034098567,
2903
+ "low": 0.9332346113687782,
2904
+ "mean": 1.153557995662505,
2905
+ "n": 5,
2906
+ "num_samples": 200
2907
+ },
2908
+ "mean": 1.153557995662505,
2909
+ "per_task": {
2910
+ "LiftPegUpright-v1": 0.8375820355789246,
2911
+ "PickCube-v1": 1.0095556863358777,
2912
+ "PullCube-v1": 1.9391989570475572,
2913
+ "PushCube-v1": 1.0238665053485367,
2914
+ "StackCube-v1": 0.9575867940016279
2915
+ }
2916
+ },
2917
+ "metric": "support_gap_fraction_to_hidden_at_8",
2918
+ "micro": {
2919
+ "confidence": 0.95,
2920
+ "high": 1.9116562543571636,
2921
+ "low": 1.121501978383282,
2922
+ "mean": 1.4795985457901433,
2923
+ "n": 50,
2924
+ "num_samples": 200
2925
+ },
2926
+ "n": 50
2927
+ },
2928
+ "total_car_to_hidden_at_8": {
2929
+ "macro_by_seed": {
2930
+ "ci": {
2931
+ "confidence": 0.95,
2932
+ "high": 1.3758006894066006,
2933
+ "low": 1.003206735187547,
2934
+ "mean": 1.2076171057325202,
2935
+ "n": 47,
2936
+ "num_samples": 200
2937
+ },
2938
+ "mean": 1.2076171057325202,
2939
+ "per_seed": {
2940
+ "1025740212": 1.7932826280593872,
2941
+ "1055157354": 1.7967484891414642,
2942
+ "1055567618": 1.1787617802619934,
2943
+ "1113268940": 1.7866801768541336,
2944
+ "1113371261": 0.0,
2945
+ "1176943004": 1.9365029335021973,
2946
+ "1189392917": 0.0,
2947
+ "1231780997": 1.7896185666322708,
2948
+ "1271718007": 0.8751109465956688,
2949
+ "1273247072": 1.9101924449205399,
2950
+ "1358421325": 1.8465763926506042,
2951
+ "1375492786": 0.24003965663723648,
2952
+ "141984615": 0.11700566027623911,
2953
+ "1475440973": 1.2767300605773926,
2954
+ "1483645336": 1.8495931923389435,
2955
+ "1522272686": 1.775540068745613,
2956
+ "1528740417": 1.7388119995594025,
2957
+ "1549324805": 1.3119094495971997,
2958
+ "1607773860": 1.1988454461097717,
2959
+ "1665739511": 0.0,
2960
+ "1747178030": 1.9230644702911377,
2961
+ "1881223807": 1.9508575201034546,
2962
+ "1919198188": 1.9435934722423553,
2963
+ "1991414747": 1.9197078198194504,
2964
+ "1997090716": 1.875569999217987,
2965
+ "2015956957": 0.42817116193473337,
2966
+ "2054899126": 0.0,
2967
+ "211245842": 1.2096436023712158,
2968
+ "2141802816": 1.8454482555389404,
2969
+ "256770874": 1.8235301077365875,
2970
+ "278008524": 1.0875415280461311,
2971
+ "416183106": 1.4655730724334717,
2972
+ "534717141": 1.910297967493534,
2973
+ "551504635": 1.2730597853660583,
2974
+ "565369392": 1.218052012224992,
2975
+ "565621826": 0.0,
2976
+ "585904130": 0.16900387903054556,
2977
+ "626486813": 0.5373998517170548,
2978
+ "636680137": 1.9735268149524927,
2979
+ "706479600": 1.1812418699264526,
2980
+ "750448296": 0.11358869262039661,
2981
+ "75070641": 1.807526096701622,
2982
+ "767747424": 0.09620045265182853,
2983
+ "77147264": 0.9106300845742226,
2984
+ "794181626": 0.0,
2985
+ "957101489": 1.7651904374361038,
2986
+ "998761162": 1.907635122537613
2987
+ }
2988
+ },
2989
+ "macro_by_task": {
2990
+ "ci": {
2991
+ "confidence": 0.95,
2992
+ "high": 1.7026664297550151,
2993
+ "low": 0.9626722220255577,
2994
+ "mean": 1.3547868940304422,
2995
+ "n": 5,
2996
+ "num_samples": 200
2997
+ },
2998
+ "mean": 1.3547868940304422,
2999
+ "per_task": {
3000
+ "LiftPegUpright-v1": 1.609697202841441,
3001
+ "PickCube-v1": 1.9101924449205399,
3002
+ "PullCube-v1": 0.5767036846784108,
3003
+ "PushCube-v1": 1.173057611172016,
3004
+ "StackCube-v1": 1.5042835265398025
3005
+ }
3006
+ },
3007
+ "metric": "total_car_to_hidden_at_8",
3008
+ "micro": {
3009
+ "confidence": 0.95,
3010
+ "high": 1.2047905542017163,
3011
+ "low": 0.7622993884682385,
3012
+ "mean": 0.9998594606430202,
3013
+ "n": 69,
3014
+ "num_samples": 200
3015
+ },
3016
+ "n": 69
3017
+ }
3018
+ },
3019
+ "num_rows": 69,
3020
+ "report_type": "ctt_generated_measured_rollout",
3021
+ "schema_version": 1,
3022
+ "source_content_hash": "67f7b4a692f7c7e71da24378dc71e3399c5d35afc14c494869eba6087b765c42",
3023
+ "source_split_hash": "867a94a1c67686c082765671a7bf8d74a9d633085e1b2619c83197c60bf58372",
3024
+ "success_summary": {
3025
+ "base_success_rate": 0.2753623188405797,
3026
+ "base_utility_mean": 0.7084005813019864,
3027
+ "hidden_chart_oracle_success_rate": 0.6666666666666666,
3028
+ "hidden_chart_oracle_utility_mean": 1.3999846732054932,
3029
+ "max_restore_error": 1.4901161193847656e-08,
3030
+ "proposal_oracle_success_gain_over_base": -0.11594202898550725,
3031
+ "proposal_oracle_success_rate": 0.15942028985507245,
3032
+ "proposal_oracle_utility_mean": 0.5035007005098505,
3033
+ "selected_success_gain_over_base": -0.15942028985507245,
3034
+ "selected_success_rate": 0.11594202898550725,
3035
+ "selected_utility_mean": 0.4006981179388105,
3036
+ "success_selector_gap": 0.043478260869565216,
3037
+ "success_support_gap": 0.5072463768115942
3038
+ },
3039
+ "target_content_hash": "1aacf9de98648fa36bef14e137eb43c7c6973170804650814dae41d09842c708",
3040
+ "target_split_hash": "9d1dc1c10868f76fa080895df0f2730c173f12b6db14932599d55c2b60859381"
3041
+ }
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/metrics_by_seed.json ADDED
@@ -0,0 +1,1647 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "1025740212": {
3
+ "any_generated_unsafe_known_at_8": 0.0,
4
+ "base_safety_label_known": 1.0,
5
+ "base_success": 0.0,
6
+ "base_unsafe_known": 0.0,
7
+ "base_utility": 0.21361251175403595,
8
+ "branch_car_at_8": 0.0009215474128723145,
9
+ "generated_safety_label_coverage_at_8": 1.0,
10
+ "generated_unsafe_rate_known_at_8": 0.0,
11
+ "hidden_chart_oracle_success_at_8": 1.0,
12
+ "hidden_chart_oracle_utility_at_8": 2.0,
13
+ "ncar_to_hidden_chart_oracle_at_8": 1.0038598175480926,
14
+ "ncar_to_proposal_oracle_at_8": 0.15426996158755663,
15
+ "outcome_ptr_at_8": 0.0,
16
+ "pairwise_causal_calibration_ece": 0.23026774639555236,
17
+ "proposal_oracle_safety_label_known_at_8": 1.0,
18
+ "proposal_oracle_success_at_8": 0.0,
19
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
20
+ "proposal_oracle_unsafe_known_at_8": 0.0,
21
+ "proposal_oracle_utility_at_8": 0.2076389193534851,
22
+ "proposal_oracle_utility_gain_over_base_at_8": -0.005973592400550842,
23
+ "selected_safety_label_known_at_8": 1.0,
24
+ "selected_success_at_8": 0.0,
25
+ "selected_success_gain_over_base_at_8": 0.0,
26
+ "selected_unsafe_known_at_8": 0.0,
27
+ "selected_utility_at_8": 0.2067173719406128,
28
+ "selected_utility_gain_over_base_at_8": -0.006895139813423157,
29
+ "selector_gap_fraction_to_hidden_at_8": 0.000515872070833396,
30
+ "selector_regret_at_8": 0.0009215474128723145,
31
+ "success_selector_gap_at_8": 0.0,
32
+ "success_support_gap_at_8": 1.0,
33
+ "success_total_car_to_hidden_at_8": 1.0,
34
+ "support_gap_at_8": 1.792361080646515,
35
+ "support_gap_fraction_to_hidden_at_8": 1.0033439510967557,
36
+ "total_car_to_hidden_at_8": 1.7932826280593872
37
+ },
38
+ "1055157354": {
39
+ "any_generated_unsafe_known_at_8": 0.0,
40
+ "base_safety_label_known": 1.0,
41
+ "base_success": 0.0,
42
+ "base_unsafe_known": 0.0,
43
+ "base_utility": 0.1879976987838745,
44
+ "branch_car_at_8": 0.014723598957061768,
45
+ "generated_safety_label_coverage_at_8": 1.0,
46
+ "generated_unsafe_rate_known_at_8": 0.0,
47
+ "hidden_chart_oracle_success_at_8": 1.0,
48
+ "hidden_chart_oracle_utility_at_8": 2.0,
49
+ "ncar_to_hidden_chart_oracle_at_8": 0.9915817866344643,
50
+ "ncar_to_proposal_oracle_at_8": 0.49115629197991395,
51
+ "outcome_ptr_at_8": 1.0,
52
+ "pairwise_causal_calibration_ece": 0.19960528924679433,
53
+ "proposal_oracle_safety_label_known_at_8": 1.0,
54
+ "proposal_oracle_success_at_8": 0.0,
55
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
56
+ "proposal_oracle_unsafe_known_at_8": 0.0,
57
+ "proposal_oracle_utility_at_8": 0.21797510981559753,
58
+ "proposal_oracle_utility_gain_over_base_at_8": 0.029977411031723022,
59
+ "selected_safety_label_known_at_8": 1.0,
60
+ "selected_success_at_8": 0.0,
61
+ "selected_success_gain_over_base_at_8": 0.0,
62
+ "selected_unsafe_known_at_8": 0.0,
63
+ "selected_utility_at_8": 0.20325151085853577,
64
+ "selected_utility_gain_over_base_at_8": 0.015253812074661255,
65
+ "selector_gap_fraction_to_hidden_at_8": 0.00812559616904461,
66
+ "selector_regret_at_8": 0.014723598957061768,
67
+ "success_selector_gap_at_8": 0.0,
68
+ "success_support_gap_at_8": 1.0,
69
+ "success_total_car_to_hidden_at_8": 1.0,
70
+ "support_gap_at_8": 1.7820248901844025,
71
+ "support_gap_fraction_to_hidden_at_8": 0.9834561959377184,
72
+ "total_car_to_hidden_at_8": 1.7967484891414642
73
+ },
74
+ "1055567618": {
75
+ "any_generated_unsafe_known_at_8": 0.0,
76
+ "base_safety_label_known": 1.0,
77
+ "base_success": 1.0,
78
+ "base_unsafe_known": 0.0,
79
+ "base_utility": 2.0,
80
+ "branch_car_at_8": 0.018837809562683105,
81
+ "generated_safety_label_coverage_at_8": 1.0,
82
+ "generated_unsafe_rate_known_at_8": 0.0,
83
+ "hidden_chart_oracle_success_at_8": 1.0,
84
+ "hidden_chart_oracle_utility_at_8": 2.0,
85
+ "ncar_to_proposal_oracle_at_8": 0.016240555308914228,
86
+ "outcome_ptr_at_8": 0.0,
87
+ "pairwise_causal_calibration_ece": 0.18844312444633177,
88
+ "proposal_oracle_safety_label_known_at_8": 1.0,
89
+ "proposal_oracle_success_at_8": 0.0,
90
+ "proposal_oracle_success_gain_over_base_at_8": -1.0,
91
+ "proposal_oracle_unsafe_known_at_8": 0.0,
92
+ "proposal_oracle_utility_at_8": 0.8400760293006897,
93
+ "proposal_oracle_utility_gain_over_base_at_8": -1.1599239706993103,
94
+ "selected_safety_label_known_at_8": 1.0,
95
+ "selected_success_at_8": 0.0,
96
+ "selected_success_gain_over_base_at_8": -1.0,
97
+ "selected_unsafe_known_at_8": 0.0,
98
+ "selected_utility_at_8": 0.8212382197380066,
99
+ "selected_utility_gain_over_base_at_8": -1.1787617802619934,
100
+ "selector_regret_at_8": 0.018837809562683105,
101
+ "success_selector_gap_at_8": 0.0,
102
+ "success_support_gap_at_8": 1.0,
103
+ "success_total_car_to_hidden_at_8": 1.0,
104
+ "support_gap_at_8": 1.1599239706993103,
105
+ "total_car_to_hidden_at_8": 1.1787617802619934
106
+ },
107
+ "1113268940": {
108
+ "any_generated_unsafe_known_at_8": 0.0,
109
+ "base_safety_label_known": 1.0,
110
+ "base_success": 0.0,
111
+ "base_unsafe_known": 0.0,
112
+ "base_utility": 0.21901588141918182,
113
+ "branch_car_at_8": 0.006647542119026184,
114
+ "generated_safety_label_coverage_at_8": 1.0,
115
+ "generated_unsafe_rate_known_at_8": 0.0,
116
+ "hidden_chart_oracle_success_at_8": 1.0,
117
+ "hidden_chart_oracle_utility_at_8": 2.0,
118
+ "ncar_to_hidden_chart_oracle_at_8": 1.0031982588625619,
119
+ "outcome_ptr_at_8": 1.0,
120
+ "pairwise_causal_calibration_ece": 0.4201805991515518,
121
+ "proposal_oracle_safety_label_known_at_8": 1.0,
122
+ "proposal_oracle_success_at_8": 0.0,
123
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
124
+ "proposal_oracle_unsafe_known_at_8": 0.0,
125
+ "proposal_oracle_utility_at_8": 0.21996736526489258,
126
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0009514838457107544,
127
+ "selected_safety_label_known_at_8": 1.0,
128
+ "selected_success_at_8": 0.0,
129
+ "selected_success_gain_over_base_at_8": 0.0,
130
+ "selected_unsafe_known_at_8": 0.0,
131
+ "selected_utility_at_8": 0.2133198231458664,
132
+ "selected_utility_gain_over_base_at_8": -0.00569605827331543,
133
+ "selector_gap_fraction_to_hidden_at_8": 0.0037325106101020683,
134
+ "selector_regret_at_8": 0.006647542119026184,
135
+ "success_selector_gap_at_8": 0.0,
136
+ "success_support_gap_at_8": 1.0,
137
+ "success_total_car_to_hidden_at_8": 1.0,
138
+ "support_gap_at_8": 1.7800326347351074,
139
+ "support_gap_fraction_to_hidden_at_8": 0.9994657538852907,
140
+ "total_car_to_hidden_at_8": 1.7866801768541336
141
+ },
142
+ "1113371261": {
143
+ "any_generated_unsafe_known_at_8": 0.0,
144
+ "base_safety_label_known": 1.0,
145
+ "base_success": 1.0,
146
+ "base_unsafe_known": 0.0,
147
+ "base_utility": 2.0,
148
+ "branch_car_at_8": 0.0,
149
+ "generated_safety_label_coverage_at_8": 1.0,
150
+ "generated_unsafe_rate_known_at_8": 0.0,
151
+ "hidden_chart_oracle_success_at_8": 1.0,
152
+ "hidden_chart_oracle_utility_at_8": 2.0,
153
+ "outcome_ptr_at_8": 0.0,
154
+ "pairwise_causal_calibration_ece": 0.10528846281262146,
155
+ "proposal_oracle_safety_label_known_at_8": 1.0,
156
+ "proposal_oracle_success_at_8": 1.0,
157
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
158
+ "proposal_oracle_unsafe_known_at_8": 0.0,
159
+ "proposal_oracle_utility_at_8": 2.0,
160
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
161
+ "selected_safety_label_known_at_8": 1.0,
162
+ "selected_success_at_8": 1.0,
163
+ "selected_success_gain_over_base_at_8": 0.0,
164
+ "selected_unsafe_known_at_8": 0.0,
165
+ "selected_utility_at_8": 2.0,
166
+ "selected_utility_gain_over_base_at_8": 0.0,
167
+ "selector_regret_at_8": 0.0,
168
+ "success_selector_gap_at_8": 0.0,
169
+ "success_support_gap_at_8": 0.0,
170
+ "success_total_car_to_hidden_at_8": 0.0,
171
+ "support_gap_at_8": 0.0,
172
+ "total_car_to_hidden_at_8": 0.0
173
+ },
174
+ "1176943004": {
175
+ "any_generated_unsafe_known_at_8": 0.0,
176
+ "base_safety_label_known": 1.0,
177
+ "base_success": 1.0,
178
+ "base_unsafe_known": 0.0,
179
+ "base_utility": 2.0,
180
+ "branch_car_at_8": 0.03303612768650055,
181
+ "generated_safety_label_coverage_at_8": 1.0,
182
+ "generated_unsafe_rate_known_at_8": 0.0,
183
+ "hidden_chart_oracle_success_at_8": 1.0,
184
+ "hidden_chart_oracle_utility_at_8": 2.0,
185
+ "ncar_to_proposal_oracle_at_8": 0.01735576759836683,
186
+ "outcome_ptr_at_8": 0.0,
187
+ "pairwise_causal_calibration_ece": 0.18726924478108561,
188
+ "proposal_oracle_safety_label_known_at_8": 1.0,
189
+ "proposal_oracle_success_at_8": 0.0,
190
+ "proposal_oracle_success_gain_over_base_at_8": -1.0,
191
+ "proposal_oracle_unsafe_known_at_8": 0.0,
192
+ "proposal_oracle_utility_at_8": 0.09653319418430328,
193
+ "proposal_oracle_utility_gain_over_base_at_8": -1.9034668058156967,
194
+ "selected_safety_label_known_at_8": 1.0,
195
+ "selected_success_at_8": 0.0,
196
+ "selected_success_gain_over_base_at_8": -1.0,
197
+ "selected_unsafe_known_at_8": 0.0,
198
+ "selected_utility_at_8": 0.06349706649780273,
199
+ "selected_utility_gain_over_base_at_8": -1.9365029335021973,
200
+ "selector_regret_at_8": 0.03303612768650055,
201
+ "success_selector_gap_at_8": 0.0,
202
+ "success_support_gap_at_8": 1.0,
203
+ "success_total_car_to_hidden_at_8": 1.0,
204
+ "support_gap_at_8": 1.9034668058156967,
205
+ "total_car_to_hidden_at_8": 1.9365029335021973
206
+ },
207
+ "1189392917": {
208
+ "any_generated_unsafe_known_at_8": 0.0,
209
+ "base_safety_label_known": 1.0,
210
+ "base_success": 1.0,
211
+ "base_unsafe_known": 0.0,
212
+ "base_utility": 2.0,
213
+ "branch_car_at_8": 0.0,
214
+ "generated_safety_label_coverage_at_8": 1.0,
215
+ "generated_unsafe_rate_known_at_8": 0.0,
216
+ "hidden_chart_oracle_success_at_8": 1.0,
217
+ "hidden_chart_oracle_utility_at_8": 2.0,
218
+ "outcome_ptr_at_8": 0.0,
219
+ "pairwise_causal_calibration_ece": 0.0,
220
+ "proposal_oracle_safety_label_known_at_8": 1.0,
221
+ "proposal_oracle_success_at_8": 1.0,
222
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
223
+ "proposal_oracle_unsafe_known_at_8": 0.0,
224
+ "proposal_oracle_utility_at_8": 2.0,
225
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
226
+ "selected_safety_label_known_at_8": 1.0,
227
+ "selected_success_at_8": 1.0,
228
+ "selected_success_gain_over_base_at_8": 0.0,
229
+ "selected_unsafe_known_at_8": 0.0,
230
+ "selected_utility_at_8": 2.0,
231
+ "selected_utility_gain_over_base_at_8": 0.0,
232
+ "selector_regret_at_8": 0.0,
233
+ "success_selector_gap_at_8": 0.0,
234
+ "success_support_gap_at_8": 0.0,
235
+ "success_total_car_to_hidden_at_8": 0.0,
236
+ "support_gap_at_8": 0.0,
237
+ "total_car_to_hidden_at_8": 0.0
238
+ },
239
+ "1231780997": {
240
+ "any_generated_unsafe_known_at_8": 0.0,
241
+ "base_safety_label_known": 1.0,
242
+ "base_success": 0.0,
243
+ "base_unsafe_known": 0.0,
244
+ "base_utility": 0.21144884824752808,
245
+ "branch_car_at_8": 0.0,
246
+ "generated_safety_label_coverage_at_8": 1.0,
247
+ "generated_unsafe_rate_known_at_8": 0.0,
248
+ "hidden_chart_oracle_success_at_8": 1.0,
249
+ "hidden_chart_oracle_utility_at_8": 2.0,
250
+ "ncar_to_hidden_chart_oracle_at_8": 1.0005967986281998,
251
+ "ncar_to_proposal_oracle_at_8": 0.0,
252
+ "outcome_ptr_at_8": 0.0,
253
+ "pairwise_causal_calibration_ece": 0.13464701141472765,
254
+ "proposal_oracle_safety_label_known_at_8": 1.0,
255
+ "proposal_oracle_success_at_8": 0.0,
256
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
257
+ "proposal_oracle_unsafe_known_at_8": 0.0,
258
+ "proposal_oracle_utility_at_8": 0.2103814333677292,
259
+ "proposal_oracle_utility_gain_over_base_at_8": -0.0010674148797988892,
260
+ "selected_safety_label_known_at_8": 1.0,
261
+ "selected_success_at_8": 0.0,
262
+ "selected_success_gain_over_base_at_8": 0.0,
263
+ "selected_unsafe_known_at_8": 0.0,
264
+ "selected_utility_at_8": 0.2103814333677292,
265
+ "selected_utility_gain_over_base_at_8": -0.0010674148797988892,
266
+ "selector_gap_fraction_to_hidden_at_8": 0.0,
267
+ "selector_regret_at_8": 0.0,
268
+ "success_selector_gap_at_8": 0.0,
269
+ "success_support_gap_at_8": 1.0,
270
+ "success_total_car_to_hidden_at_8": 1.0,
271
+ "support_gap_at_8": 1.7896185666322708,
272
+ "support_gap_fraction_to_hidden_at_8": 1.0005968042226543,
273
+ "total_car_to_hidden_at_8": 1.7896185666322708
274
+ },
275
+ "1271718007": {
276
+ "any_generated_unsafe_known_at_8": 0.0,
277
+ "base_safety_label_known": 1.0,
278
+ "base_success": 0.5,
279
+ "base_unsafe_known": 0.0,
280
+ "base_utility": 1.10479174554348,
281
+ "branch_car_at_8": 0.0,
282
+ "generated_safety_label_coverage_at_8": 1.0,
283
+ "generated_unsafe_rate_known_at_8": 0.0,
284
+ "hidden_chart_oracle_success_at_8": 1.0,
285
+ "hidden_chart_oracle_utility_at_8": 2.0,
286
+ "ncar_to_hidden_chart_oracle_at_8": 0.977550125740515,
287
+ "ncar_to_proposal_oracle_at_8": 0.0,
288
+ "outcome_ptr_at_8": 0.5,
289
+ "pairwise_causal_calibration_ece": 0.07020098827281387,
290
+ "proposal_oracle_safety_label_known_at_8": 1.0,
291
+ "proposal_oracle_success_at_8": 0.5,
292
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
293
+ "proposal_oracle_unsafe_known_at_8": 0.0,
294
+ "proposal_oracle_utility_at_8": 1.1248890534043312,
295
+ "proposal_oracle_utility_gain_over_base_at_8": 0.020097307860851288,
296
+ "selected_safety_label_known_at_8": 1.0,
297
+ "selected_success_at_8": 0.5,
298
+ "selected_success_gain_over_base_at_8": 0.0,
299
+ "selected_unsafe_known_at_8": 0.0,
300
+ "selected_utility_at_8": 1.1248890534043312,
301
+ "selected_utility_gain_over_base_at_8": 0.020097307860851288,
302
+ "selector_gap_fraction_to_hidden_at_8": 0.0,
303
+ "selector_regret_at_8": 0.0,
304
+ "success_selector_gap_at_8": 0.0,
305
+ "success_support_gap_at_8": 0.5,
306
+ "success_total_car_to_hidden_at_8": 0.5,
307
+ "support_gap_at_8": 0.8751109465956688,
308
+ "support_gap_fraction_to_hidden_at_8": 0.9775501312004184,
309
+ "total_car_to_hidden_at_8": 0.8751109465956688
310
+ },
311
+ "1273247072": {
312
+ "any_generated_unsafe_known_at_8": 0.0,
313
+ "base_safety_label_known": 1.0,
314
+ "base_success": 0.0,
315
+ "base_unsafe_known": 0.0,
316
+ "base_utility": 0.11945180594921112,
317
+ "branch_car_at_8": 0.011674322187900543,
318
+ "generated_safety_label_coverage_at_8": 1.0,
319
+ "generated_unsafe_rate_known_at_8": 0.0,
320
+ "hidden_chart_oracle_success_at_8": 1.0,
321
+ "hidden_chart_oracle_utility_at_8": 2.0,
322
+ "ncar_to_hidden_chart_oracle_at_8": 1.0157636165911068,
323
+ "ncar_to_proposal_oracle_at_8": 0.649658432039702,
324
+ "outcome_ptr_at_8": 0.0,
325
+ "pairwise_causal_calibration_ece": 0.11483608299984571,
326
+ "proposal_oracle_safety_label_known_at_8": 1.0,
327
+ "proposal_oracle_success_at_8": 0.0,
328
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
329
+ "proposal_oracle_unsafe_known_at_8": 0.0,
330
+ "proposal_oracle_utility_at_8": 0.10148187726736069,
331
+ "proposal_oracle_utility_gain_over_base_at_8": -0.017969928681850433,
332
+ "selected_safety_label_known_at_8": 1.0,
333
+ "selected_success_at_8": 0.0,
334
+ "selected_success_gain_over_base_at_8": 0.0,
335
+ "selected_unsafe_known_at_8": 0.0,
336
+ "selected_utility_at_8": 0.08980755507946014,
337
+ "selected_utility_gain_over_base_at_8": -0.029644250869750977,
338
+ "selector_gap_fraction_to_hidden_at_8": 0.006207935656652067,
339
+ "selector_regret_at_8": 0.011674322187900543,
340
+ "success_selector_gap_at_8": 0.0,
341
+ "success_support_gap_at_8": 1.0,
342
+ "success_total_car_to_hidden_at_8": 1.0,
343
+ "support_gap_at_8": 1.8985181227326393,
344
+ "support_gap_fraction_to_hidden_at_8": 1.0095556863358777,
345
+ "total_car_to_hidden_at_8": 1.9101924449205399
346
+ },
347
+ "1358421325": {
348
+ "any_generated_unsafe_known_at_8": 0.0,
349
+ "base_safety_label_known": 1.0,
350
+ "base_success": 1.0,
351
+ "base_unsafe_known": 0.0,
352
+ "base_utility": 2.0,
353
+ "branch_car_at_8": 0.6867138147354126,
354
+ "generated_safety_label_coverage_at_8": 1.0,
355
+ "generated_unsafe_rate_known_at_8": 0.0,
356
+ "hidden_chart_oracle_success_at_8": 1.0,
357
+ "hidden_chart_oracle_utility_at_8": 2.0,
358
+ "ncar_to_proposal_oracle_at_8": 0.5920648030985759,
359
+ "outcome_ptr_at_8": 0.0,
360
+ "pairwise_causal_calibration_ece": 0.24959422175257154,
361
+ "proposal_oracle_safety_label_known_at_8": 1.0,
362
+ "proposal_oracle_success_at_8": 0.0,
363
+ "proposal_oracle_success_gain_over_base_at_8": -1.0,
364
+ "proposal_oracle_unsafe_known_at_8": 0.0,
365
+ "proposal_oracle_utility_at_8": 0.8401374220848083,
366
+ "proposal_oracle_utility_gain_over_base_at_8": -1.1598625779151917,
367
+ "selected_safety_label_known_at_8": 1.0,
368
+ "selected_success_at_8": 0.0,
369
+ "selected_success_gain_over_base_at_8": -1.0,
370
+ "selected_unsafe_known_at_8": 0.0,
371
+ "selected_utility_at_8": 0.15342360734939575,
372
+ "selected_utility_gain_over_base_at_8": -1.8465763926506042,
373
+ "selector_regret_at_8": 0.6867138147354126,
374
+ "success_selector_gap_at_8": 0.0,
375
+ "success_support_gap_at_8": 1.0,
376
+ "success_total_car_to_hidden_at_8": 1.0,
377
+ "support_gap_at_8": 1.1598625779151917,
378
+ "total_car_to_hidden_at_8": 1.8465763926506042
379
+ },
380
+ "1375492786": {
381
+ "any_generated_unsafe_known_at_8": 1.0,
382
+ "base_safety_label_known": 1.0,
383
+ "base_success": 0.0,
384
+ "base_unsafe_known": 0.0,
385
+ "base_utility": 0.18323469161987305,
386
+ "branch_car_at_8": 0.0017340581398457289,
387
+ "generated_safety_label_coverage_at_8": 1.0,
388
+ "generated_unsafe_rate_known_at_8": 1.0,
389
+ "hidden_chart_oracle_success_at_8": 0.0,
390
+ "hidden_chart_oracle_utility_at_8": 0.24345096945762634,
391
+ "ncar_to_hidden_chart_oracle_at_8": 3.9862911723153553,
392
+ "ncar_to_proposal_oracle_at_8": 0.009737013067671323,
393
+ "outcome_ptr_at_8": 0.0,
394
+ "pairwise_causal_calibration_ece": 0.13157026218508128,
395
+ "proposal_oracle_safety_label_known_at_8": 1.0,
396
+ "proposal_oracle_success_at_8": 0.0,
397
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
398
+ "proposal_oracle_unsafe_known_at_8": 1.0,
399
+ "proposal_oracle_utility_at_8": 0.005145370960235596,
400
+ "proposal_oracle_utility_gain_over_base_at_8": -0.17808932065963745,
401
+ "selected_safety_label_known_at_8": 1.0,
402
+ "selected_success_at_8": 0.0,
403
+ "selected_success_gain_over_base_at_8": 0.0,
404
+ "selected_unsafe_known_at_8": 1.0,
405
+ "selected_utility_at_8": 0.003411312820389867,
406
+ "selected_utility_gain_over_base_at_8": -0.17982337879948318,
407
+ "selector_gap_fraction_to_hidden_at_8": 0.028797165851366206,
408
+ "selector_regret_at_8": 0.0017340581398457289,
409
+ "success_selector_gap_at_8": 0.0,
410
+ "success_support_gap_at_8": 0.0,
411
+ "success_total_car_to_hidden_at_8": 0.0,
412
+ "support_gap_at_8": 0.23830559849739075,
413
+ "support_gap_fraction_to_hidden_at_8": 3.957494668459602,
414
+ "total_car_to_hidden_at_8": 0.24003965663723648
415
+ },
416
+ "141984615": {
417
+ "any_generated_unsafe_known_at_8": 0.3333333333333333,
418
+ "base_safety_label_known": 1.0,
419
+ "base_success": 0.0,
420
+ "base_unsafe_known": 0.0,
421
+ "base_utility": 0.12483831991751988,
422
+ "branch_car_at_8": 0.017901106815164287,
423
+ "generated_safety_label_coverage_at_8": 1.0,
424
+ "generated_unsafe_rate_known_at_8": 0.3333333333333333,
425
+ "hidden_chart_oracle_success_at_8": 0.0,
426
+ "hidden_chart_oracle_utility_at_8": 0.20044784247875214,
427
+ "ncar_to_hidden_chart_oracle_at_8": 3.2387927981827365,
428
+ "ncar_to_proposal_oracle_at_8": 0.20565346679569363,
429
+ "outcome_ptr_at_8": 0.3333333333333333,
430
+ "pairwise_causal_calibration_ece": 0.17017158124656792,
431
+ "proposal_oracle_safety_label_known_at_8": 1.0,
432
+ "proposal_oracle_success_at_8": 0.0,
433
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
434
+ "proposal_oracle_unsafe_known_at_8": 0.3333333333333333,
435
+ "proposal_oracle_utility_at_8": 0.10611975689729054,
436
+ "proposal_oracle_utility_gain_over_base_at_8": -0.01871856302022934,
437
+ "selected_safety_label_known_at_8": 1.0,
438
+ "selected_success_at_8": 0.0,
439
+ "selected_success_gain_over_base_at_8": 0.0,
440
+ "selected_unsafe_known_at_8": 0.3333333333333333,
441
+ "selected_utility_at_8": 0.08821865008212626,
442
+ "selected_utility_gain_over_base_at_8": -0.03661966983539363,
443
+ "selector_gap_fraction_to_hidden_at_8": 0.6932408439097882,
444
+ "selector_regret_at_8": 0.017901106815164287,
445
+ "success_selector_gap_at_8": 0.0,
446
+ "success_support_gap_at_8": 0.0,
447
+ "success_total_car_to_hidden_at_8": 0.0,
448
+ "support_gap_at_8": 0.11338664591312408,
449
+ "support_gap_fraction_to_hidden_at_8": 2.951807933296765,
450
+ "total_car_to_hidden_at_8": 0.11700566027623911
451
+ },
452
+ "1475440973": {
453
+ "any_generated_unsafe_known_at_8": 0.0,
454
+ "base_safety_label_known": 1.0,
455
+ "base_success": 0.0,
456
+ "base_unsafe_known": 0.0,
457
+ "base_utility": 0.7229747772216797,
458
+ "branch_car_at_8": 0.004936397075653076,
459
+ "generated_safety_label_coverage_at_8": 1.0,
460
+ "generated_unsafe_rate_known_at_8": 0.0,
461
+ "hidden_chart_oracle_success_at_8": 1.0,
462
+ "hidden_chart_oracle_utility_at_8": 2.0,
463
+ "ncar_to_hidden_chart_oracle_at_8": 0.9997688595390668,
464
+ "ncar_to_proposal_oracle_at_8": 0.943578650052657,
465
+ "outcome_ptr_at_8": 1.0,
466
+ "pairwise_causal_calibration_ece": 0.11407560353818716,
467
+ "proposal_oracle_safety_label_known_at_8": 1.0,
468
+ "proposal_oracle_success_at_8": 0.0,
469
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
470
+ "proposal_oracle_unsafe_known_at_8": 0.0,
471
+ "proposal_oracle_utility_at_8": 0.7282063364982605,
472
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0052315592765808105,
473
+ "selected_safety_label_known_at_8": 1.0,
474
+ "selected_success_at_8": 0.0,
475
+ "selected_success_gain_over_base_at_8": 0.0,
476
+ "selected_unsafe_known_at_8": 0.0,
477
+ "selected_utility_at_8": 0.7232699394226074,
478
+ "selected_utility_gain_over_base_at_8": 0.0002951622009277344,
479
+ "selector_gap_fraction_to_hidden_at_8": 0.0038655439122129136,
480
+ "selector_regret_at_8": 0.004936397075653076,
481
+ "success_selector_gap_at_8": 0.0,
482
+ "success_support_gap_at_8": 1.0,
483
+ "success_total_car_to_hidden_at_8": 1.0,
484
+ "support_gap_at_8": 1.2717936635017395,
485
+ "support_gap_fraction_to_hidden_at_8": 0.9959033234557427,
486
+ "total_car_to_hidden_at_8": 1.2767300605773926
487
+ },
488
+ "1483645336": {
489
+ "any_generated_unsafe_known_at_8": 0.0,
490
+ "base_safety_label_known": 1.0,
491
+ "base_success": 1.0,
492
+ "base_unsafe_known": 0.0,
493
+ "base_utility": 2.0,
494
+ "branch_car_at_8": 1.8495931923389435,
495
+ "generated_safety_label_coverage_at_8": 1.0,
496
+ "generated_unsafe_rate_known_at_8": 0.0,
497
+ "hidden_chart_oracle_success_at_8": 1.0,
498
+ "hidden_chart_oracle_utility_at_8": 2.0,
499
+ "outcome_ptr_at_8": 0.0,
500
+ "pairwise_causal_calibration_ece": 0.5238458588992289,
501
+ "proposal_oracle_safety_label_known_at_8": 1.0,
502
+ "proposal_oracle_success_at_8": 1.0,
503
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
504
+ "proposal_oracle_unsafe_known_at_8": 0.0,
505
+ "proposal_oracle_utility_at_8": 2.0,
506
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
507
+ "selected_safety_label_known_at_8": 1.0,
508
+ "selected_success_at_8": 0.0,
509
+ "selected_success_gain_over_base_at_8": -1.0,
510
+ "selected_unsafe_known_at_8": 0.0,
511
+ "selected_utility_at_8": 0.15040680766105652,
512
+ "selected_utility_gain_over_base_at_8": -1.8495931923389435,
513
+ "selector_regret_at_8": 1.8495931923389435,
514
+ "success_selector_gap_at_8": 1.0,
515
+ "success_support_gap_at_8": 0.0,
516
+ "success_total_car_to_hidden_at_8": 1.0,
517
+ "support_gap_at_8": 0.0,
518
+ "total_car_to_hidden_at_8": 1.8495931923389435
519
+ },
520
+ "1522272686": {
521
+ "any_generated_unsafe_known_at_8": 0.0,
522
+ "base_safety_label_known": 1.0,
523
+ "base_success": 0.0,
524
+ "base_unsafe_known": 0.0,
525
+ "base_utility": 0.5529977083206177,
526
+ "branch_car_at_8": 0.3386564999818802,
527
+ "generated_safety_label_coverage_at_8": 1.0,
528
+ "generated_unsafe_rate_known_at_8": 0.0,
529
+ "hidden_chart_oracle_success_at_8": 1.0,
530
+ "hidden_chart_oracle_utility_at_8": 2.0,
531
+ "ncar_to_hidden_chart_oracle_at_8": 1.2270471627342485,
532
+ "ncar_to_proposal_oracle_at_8": 33.46827145305134,
533
+ "outcome_ptr_at_8": 1.0,
534
+ "pairwise_causal_calibration_ece": 0.0746614940687875,
535
+ "proposal_oracle_safety_label_known_at_8": 1.0,
536
+ "proposal_oracle_success_at_8": 0.0,
537
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
538
+ "proposal_oracle_unsafe_known_at_8": 0.0,
539
+ "proposal_oracle_utility_at_8": 0.5631164312362671,
540
+ "proposal_oracle_utility_gain_over_base_at_8": 0.010118722915649414,
541
+ "selected_safety_label_known_at_8": 1.0,
542
+ "selected_success_at_8": 0.0,
543
+ "selected_success_gain_over_base_at_8": 0.0,
544
+ "selected_unsafe_known_at_8": 0.0,
545
+ "selected_utility_at_8": 0.2244599312543869,
546
+ "selected_utility_gain_over_base_at_8": -0.3285377770662308,
547
+ "selector_gap_fraction_to_hidden_at_8": 0.23404005780034906,
548
+ "selector_regret_at_8": 0.3386564999818802,
549
+ "success_selector_gap_at_8": 0.0,
550
+ "success_support_gap_at_8": 1.0,
551
+ "success_total_car_to_hidden_at_8": 1.0,
552
+ "support_gap_at_8": 1.436883568763733,
553
+ "support_gap_fraction_to_hidden_at_8": 0.9930071134138249,
554
+ "total_car_to_hidden_at_8": 1.775540068745613
555
+ },
556
+ "1528740417": {
557
+ "any_generated_unsafe_known_at_8": 0.0,
558
+ "base_safety_label_known": 1.0,
559
+ "base_success": 0.0,
560
+ "base_unsafe_known": 0.0,
561
+ "base_utility": 0.32647061347961426,
562
+ "branch_car_at_8": 0.02146562933921814,
563
+ "generated_safety_label_coverage_at_8": 1.0,
564
+ "generated_unsafe_rate_known_at_8": 0.0,
565
+ "hidden_chart_oracle_success_at_8": 1.0,
566
+ "hidden_chart_oracle_utility_at_8": 2.0,
567
+ "ncar_to_hidden_chart_oracle_at_8": 1.0390089371449063,
568
+ "ncar_to_proposal_oracle_at_8": 0.48989279105464534,
569
+ "outcome_ptr_at_8": 0.0,
570
+ "pairwise_causal_calibration_ece": 0.13131517178411375,
571
+ "proposal_oracle_safety_label_known_at_8": 1.0,
572
+ "proposal_oracle_success_at_8": 0.0,
573
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
574
+ "proposal_oracle_unsafe_known_at_8": 0.0,
575
+ "proposal_oracle_utility_at_8": 0.2826536297798157,
576
+ "proposal_oracle_utility_gain_over_base_at_8": -0.043816983699798584,
577
+ "selected_safety_label_known_at_8": 1.0,
578
+ "selected_success_at_8": 0.0,
579
+ "selected_success_gain_over_base_at_8": 0.0,
580
+ "selected_unsafe_known_at_8": 0.0,
581
+ "selected_utility_at_8": 0.26118800044059753,
582
+ "selected_utility_gain_over_base_at_8": -0.06528261303901672,
583
+ "selector_gap_fraction_to_hidden_at_8": 0.01282656254029075,
584
+ "selector_regret_at_8": 0.02146562933921814,
585
+ "success_selector_gap_at_8": 0.0,
586
+ "success_support_gap_at_8": 1.0,
587
+ "success_total_car_to_hidden_at_8": 1.0,
588
+ "support_gap_at_8": 1.7173463702201843,
589
+ "support_gap_fraction_to_hidden_at_8": 1.026182380813105,
590
+ "total_car_to_hidden_at_8": 1.7388119995594025
591
+ },
592
+ "1549324805": {
593
+ "any_generated_unsafe_known_at_8": 1.0,
594
+ "base_safety_label_known": 1.0,
595
+ "base_success": 0.3333333333333333,
596
+ "base_unsafe_known": 0.0,
597
+ "base_utility": 0.7632389987508456,
598
+ "branch_car_at_8": 0.6191183490057787,
599
+ "generated_safety_label_coverage_at_8": 1.0,
600
+ "generated_unsafe_rate_known_at_8": 0.75,
601
+ "hidden_chart_oracle_success_at_8": 0.6666666666666666,
602
+ "hidden_chart_oracle_utility_at_8": 1.4027505218982697,
603
+ "ncar_to_hidden_chart_oracle_at_8": 2.43115209329291,
604
+ "ncar_to_proposal_oracle_at_8": 0.21532264167370982,
605
+ "outcome_ptr_at_8": 0.0,
606
+ "pairwise_causal_calibration_ece": 0.3542532249847814,
607
+ "proposal_oracle_safety_label_known_at_8": 1.0,
608
+ "proposal_oracle_success_at_8": 0.3333333333333333,
609
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
610
+ "proposal_oracle_unsafe_known_at_8": 0.6666666666666666,
611
+ "proposal_oracle_utility_at_8": 0.7099594213068485,
612
+ "proposal_oracle_utility_gain_over_base_at_8": -0.05327957744399706,
613
+ "selected_safety_label_known_at_8": 1.0,
614
+ "selected_success_at_8": 0.0,
615
+ "selected_success_gain_over_base_at_8": -0.3333333333333333,
616
+ "selected_unsafe_known_at_8": 0.3333333333333333,
617
+ "selected_utility_at_8": 0.0908410723010699,
618
+ "selected_utility_gain_over_base_at_8": -0.6723979264497757,
619
+ "selector_gap_fraction_to_hidden_at_8": 0.30888217864267864,
620
+ "selector_regret_at_8": 0.6191183490057787,
621
+ "success_selector_gap_at_8": 0.3333333333333333,
622
+ "success_support_gap_at_8": 0.3333333333333333,
623
+ "success_total_car_to_hidden_at_8": 0.6666666666666666,
624
+ "support_gap_at_8": 0.6927911005914211,
625
+ "support_gap_fraction_to_hidden_at_8": 2.1222703900724977,
626
+ "total_car_to_hidden_at_8": 1.3119094495971997
627
+ },
628
+ "1607773860": {
629
+ "any_generated_unsafe_known_at_8": 0.0,
630
+ "base_safety_label_known": 1.0,
631
+ "base_success": 0.0,
632
+ "base_unsafe_known": 0.0,
633
+ "base_utility": 0.8032726645469666,
634
+ "branch_car_at_8": 0.011432766914367676,
635
+ "generated_safety_label_coverage_at_8": 1.0,
636
+ "generated_unsafe_rate_known_at_8": 0.0,
637
+ "hidden_chart_oracle_success_at_8": 1.0,
638
+ "hidden_chart_oracle_utility_at_8": 2.0,
639
+ "ncar_to_hidden_chart_oracle_at_8": 1.0017699108028124,
640
+ "ncar_to_proposal_oracle_at_8": 1.2273941543534532,
641
+ "outcome_ptr_at_8": 1.0,
642
+ "pairwise_causal_calibration_ece": 0.251291053201395,
643
+ "proposal_oracle_safety_label_known_at_8": 1.0,
644
+ "proposal_oracle_success_at_8": 0.0,
645
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
646
+ "proposal_oracle_unsafe_known_at_8": 0.0,
647
+ "proposal_oracle_utility_at_8": 0.812587320804596,
648
+ "proposal_oracle_utility_gain_over_base_at_8": 0.009314656257629395,
649
+ "selected_safety_label_known_at_8": 1.0,
650
+ "selected_success_at_8": 0.0,
651
+ "selected_success_gain_over_base_at_8": 0.0,
652
+ "selected_unsafe_known_at_8": 0.0,
653
+ "selected_utility_at_8": 0.8011545538902283,
654
+ "selected_utility_gain_over_base_at_8": -0.0021181106567382812,
655
+ "selector_gap_fraction_to_hidden_at_8": 0.009553359880460726,
656
+ "selector_regret_at_8": 0.011432766914367676,
657
+ "success_selector_gap_at_8": 0.0,
658
+ "success_support_gap_at_8": 1.0,
659
+ "success_total_car_to_hidden_at_8": 1.0,
660
+ "support_gap_at_8": 1.187412679195404,
661
+ "support_gap_fraction_to_hidden_at_8": 0.9922165592932636,
662
+ "total_car_to_hidden_at_8": 1.1988454461097717
663
+ },
664
+ "1665739511": {
665
+ "any_generated_unsafe_known_at_8": 0.0,
666
+ "base_safety_label_known": 1.0,
667
+ "base_success": 1.0,
668
+ "base_unsafe_known": 0.0,
669
+ "base_utility": 2.0,
670
+ "branch_car_at_8": 0.0,
671
+ "generated_safety_label_coverage_at_8": 1.0,
672
+ "generated_unsafe_rate_known_at_8": 0.0,
673
+ "hidden_chart_oracle_success_at_8": 1.0,
674
+ "hidden_chart_oracle_utility_at_8": 2.0,
675
+ "outcome_ptr_at_8": 0.0,
676
+ "pairwise_causal_calibration_ece": 0.0,
677
+ "proposal_oracle_safety_label_known_at_8": 1.0,
678
+ "proposal_oracle_success_at_8": 1.0,
679
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
680
+ "proposal_oracle_unsafe_known_at_8": 0.0,
681
+ "proposal_oracle_utility_at_8": 2.0,
682
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
683
+ "selected_safety_label_known_at_8": 1.0,
684
+ "selected_success_at_8": 1.0,
685
+ "selected_success_gain_over_base_at_8": 0.0,
686
+ "selected_unsafe_known_at_8": 0.0,
687
+ "selected_utility_at_8": 2.0,
688
+ "selected_utility_gain_over_base_at_8": 0.0,
689
+ "selector_regret_at_8": 0.0,
690
+ "success_selector_gap_at_8": 0.0,
691
+ "success_support_gap_at_8": 0.0,
692
+ "success_total_car_to_hidden_at_8": 0.0,
693
+ "support_gap_at_8": 0.0,
694
+ "total_car_to_hidden_at_8": 0.0
695
+ },
696
+ "1747178030": {
697
+ "any_generated_unsafe_known_at_8": 1.0,
698
+ "base_safety_label_known": 1.0,
699
+ "base_success": 1.0,
700
+ "base_unsafe_known": 0.0,
701
+ "base_utility": 2.0,
702
+ "branch_car_at_8": 0.01012641191482544,
703
+ "generated_safety_label_coverage_at_8": 1.0,
704
+ "generated_unsafe_rate_known_at_8": 1.0,
705
+ "hidden_chart_oracle_success_at_8": 1.0,
706
+ "hidden_chart_oracle_utility_at_8": 2.0,
707
+ "ncar_to_proposal_oracle_at_8": 0.0052936433657889735,
708
+ "outcome_ptr_at_8": 0.0,
709
+ "pairwise_causal_calibration_ece": 0.16974044575049851,
710
+ "proposal_oracle_safety_label_known_at_8": 1.0,
711
+ "proposal_oracle_success_at_8": 0.0,
712
+ "proposal_oracle_success_gain_over_base_at_8": -1.0,
713
+ "proposal_oracle_unsafe_known_at_8": 1.0,
714
+ "proposal_oracle_utility_at_8": 0.08706194162368774,
715
+ "proposal_oracle_utility_gain_over_base_at_8": -1.9129380583763123,
716
+ "selected_safety_label_known_at_8": 1.0,
717
+ "selected_success_at_8": 0.0,
718
+ "selected_success_gain_over_base_at_8": -1.0,
719
+ "selected_unsafe_known_at_8": 1.0,
720
+ "selected_utility_at_8": 0.0769355297088623,
721
+ "selected_utility_gain_over_base_at_8": -1.9230644702911377,
722
+ "selector_regret_at_8": 0.01012641191482544,
723
+ "success_selector_gap_at_8": 0.0,
724
+ "success_support_gap_at_8": 1.0,
725
+ "success_total_car_to_hidden_at_8": 1.0,
726
+ "support_gap_at_8": 1.9129380583763123,
727
+ "total_car_to_hidden_at_8": 1.9230644702911377
728
+ },
729
+ "1881223807": {
730
+ "any_generated_unsafe_known_at_8": 0.0,
731
+ "base_safety_label_known": 1.0,
732
+ "base_success": 0.0,
733
+ "base_unsafe_known": 0.0,
734
+ "base_utility": 0.05987149477005005,
735
+ "branch_car_at_8": 0.0018659532070159912,
736
+ "generated_safety_label_coverage_at_8": 1.0,
737
+ "generated_unsafe_rate_known_at_8": 0.0,
738
+ "hidden_chart_oracle_success_at_8": 1.0,
739
+ "hidden_chart_oracle_utility_at_8": 2.0,
740
+ "ncar_to_hidden_chart_oracle_at_8": 1.0055300485453837,
741
+ "ncar_to_proposal_oracle_at_8": 0.21053121053631743,
742
+ "outcome_ptr_at_8": 0.0,
743
+ "pairwise_causal_calibration_ece": 0.07419200542238083,
744
+ "proposal_oracle_safety_label_known_at_8": 1.0,
745
+ "proposal_oracle_success_at_8": 0.0,
746
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
747
+ "proposal_oracle_unsafe_known_at_8": 0.0,
748
+ "proposal_oracle_utility_at_8": 0.0510084331035614,
749
+ "proposal_oracle_utility_gain_over_base_at_8": -0.008863061666488647,
750
+ "selected_safety_label_known_at_8": 1.0,
751
+ "selected_success_at_8": 0.0,
752
+ "selected_success_gain_over_base_at_8": 0.0,
753
+ "selected_unsafe_known_at_8": 0.0,
754
+ "selected_utility_at_8": 0.04914247989654541,
755
+ "selected_utility_gain_over_base_at_8": -0.010729014873504639,
756
+ "selector_gap_fraction_to_hidden_at_8": 0.0009617678426897979,
757
+ "selector_regret_at_8": 0.0018659532070159912,
758
+ "success_selector_gap_at_8": 0.0,
759
+ "success_support_gap_at_8": 1.0,
760
+ "success_total_car_to_hidden_at_8": 1.0,
761
+ "support_gap_at_8": 1.9489915668964386,
762
+ "support_gap_fraction_to_hidden_at_8": 1.0045682858854952,
763
+ "total_car_to_hidden_at_8": 1.9508575201034546
764
+ },
765
+ "1919198188": {
766
+ "any_generated_unsafe_known_at_8": 0.0,
767
+ "base_safety_label_known": 1.0,
768
+ "base_success": 0.0,
769
+ "base_unsafe_known": 0.0,
770
+ "base_utility": 0.06612768769264221,
771
+ "branch_car_at_8": 0.0067754387855529785,
772
+ "generated_safety_label_coverage_at_8": 1.0,
773
+ "generated_unsafe_rate_known_at_8": 0.0,
774
+ "hidden_chart_oracle_success_at_8": 1.0,
775
+ "hidden_chart_oracle_utility_at_8": 2.0,
776
+ "ncar_to_hidden_chart_oracle_at_8": 1.0050267795980445,
777
+ "ncar_to_proposal_oracle_at_8": 2.3000872930411496,
778
+ "outcome_ptr_at_8": 0.0,
779
+ "pairwise_causal_calibration_ece": 0.16604478136348194,
780
+ "proposal_oracle_safety_label_known_at_8": 1.0,
781
+ "proposal_oracle_success_at_8": 0.0,
782
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
783
+ "proposal_oracle_unsafe_known_at_8": 0.0,
784
+ "proposal_oracle_utility_at_8": 0.06318196654319763,
785
+ "proposal_oracle_utility_gain_over_base_at_8": -0.00294572114944458,
786
+ "selected_safety_label_known_at_8": 1.0,
787
+ "selected_success_at_8": 0.0,
788
+ "selected_success_gain_over_base_at_8": 0.0,
789
+ "selected_unsafe_known_at_8": 0.0,
790
+ "selected_utility_at_8": 0.05640652775764465,
791
+ "selected_utility_gain_over_base_at_8": -0.009721159934997559,
792
+ "selector_gap_fraction_to_hidden_at_8": 0.003503560572450107,
793
+ "selector_regret_at_8": 0.0067754387855529785,
794
+ "success_selector_gap_at_8": 0.0,
795
+ "success_support_gap_at_8": 1.0,
796
+ "success_total_car_to_hidden_at_8": 1.0,
797
+ "support_gap_at_8": 1.9368180334568024,
798
+ "support_gap_fraction_to_hidden_at_8": 1.00152322422256,
799
+ "total_car_to_hidden_at_8": 1.9435934722423553
800
+ },
801
+ "1991414747": {
802
+ "any_generated_unsafe_known_at_8": 0.0,
803
+ "base_safety_label_known": 1.0,
804
+ "base_success": 1.0,
805
+ "base_unsafe_known": 0.0,
806
+ "base_utility": 2.0,
807
+ "branch_car_at_8": 0.004616528749465942,
808
+ "generated_safety_label_coverage_at_8": 1.0,
809
+ "generated_unsafe_rate_known_at_8": 0.0,
810
+ "hidden_chart_oracle_success_at_8": 1.0,
811
+ "hidden_chart_oracle_utility_at_8": 2.0,
812
+ "ncar_to_proposal_oracle_at_8": 0.002410605043679449,
813
+ "outcome_ptr_at_8": 0.0,
814
+ "pairwise_causal_calibration_ece": 0.3416902568888008,
815
+ "proposal_oracle_safety_label_known_at_8": 1.0,
816
+ "proposal_oracle_success_at_8": 0.0,
817
+ "proposal_oracle_success_gain_over_base_at_8": -1.0,
818
+ "proposal_oracle_unsafe_known_at_8": 0.0,
819
+ "proposal_oracle_utility_at_8": 0.08490870893001556,
820
+ "proposal_oracle_utility_gain_over_base_at_8": -1.9150912910699844,
821
+ "selected_safety_label_known_at_8": 1.0,
822
+ "selected_success_at_8": 0.0,
823
+ "selected_success_gain_over_base_at_8": -1.0,
824
+ "selected_unsafe_known_at_8": 0.0,
825
+ "selected_utility_at_8": 0.08029218018054962,
826
+ "selected_utility_gain_over_base_at_8": -1.9197078198194504,
827
+ "selector_regret_at_8": 0.004616528749465942,
828
+ "success_selector_gap_at_8": 0.0,
829
+ "success_support_gap_at_8": 1.0,
830
+ "success_total_car_to_hidden_at_8": 1.0,
831
+ "support_gap_at_8": 1.9150912910699844,
832
+ "total_car_to_hidden_at_8": 1.9197078198194504
833
+ },
834
+ "1997090716": {
835
+ "any_generated_unsafe_known_at_8": 0.0,
836
+ "base_safety_label_known": 1.0,
837
+ "base_success": 1.0,
838
+ "base_unsafe_known": 0.0,
839
+ "base_utility": 2.0,
840
+ "branch_car_at_8": 0.007344692945480347,
841
+ "generated_safety_label_coverage_at_8": 1.0,
842
+ "generated_unsafe_rate_known_at_8": 0.0,
843
+ "hidden_chart_oracle_success_at_8": 1.0,
844
+ "hidden_chart_oracle_utility_at_8": 2.0,
845
+ "ncar_to_proposal_oracle_at_8": 0.003931374273492085,
846
+ "outcome_ptr_at_8": 0.0,
847
+ "pairwise_causal_calibration_ece": 0.300657409601606,
848
+ "proposal_oracle_safety_label_known_at_8": 1.0,
849
+ "proposal_oracle_success_at_8": 0.0,
850
+ "proposal_oracle_success_gain_over_base_at_8": -1.0,
851
+ "proposal_oracle_unsafe_known_at_8": 0.0,
852
+ "proposal_oracle_utility_at_8": 0.1317746937274933,
853
+ "proposal_oracle_utility_gain_over_base_at_8": -1.8682253062725067,
854
+ "selected_safety_label_known_at_8": 1.0,
855
+ "selected_success_at_8": 0.0,
856
+ "selected_success_gain_over_base_at_8": -1.0,
857
+ "selected_unsafe_known_at_8": 0.0,
858
+ "selected_utility_at_8": 0.12443000078201294,
859
+ "selected_utility_gain_over_base_at_8": -1.875569999217987,
860
+ "selector_regret_at_8": 0.007344692945480347,
861
+ "success_selector_gap_at_8": 0.0,
862
+ "success_support_gap_at_8": 1.0,
863
+ "success_total_car_to_hidden_at_8": 1.0,
864
+ "support_gap_at_8": 1.8682253062725067,
865
+ "total_car_to_hidden_at_8": 1.875569999217987
866
+ },
867
+ "2015956957": {
868
+ "any_generated_unsafe_known_at_8": 0.8,
869
+ "base_safety_label_known": 1.0,
870
+ "base_success": 0.2,
871
+ "base_unsafe_known": 0.0,
872
+ "base_utility": 0.4729121312499046,
873
+ "branch_car_at_8": 0.01736527867615223,
874
+ "generated_safety_label_coverage_at_8": 1.0,
875
+ "generated_unsafe_rate_known_at_8": 0.525,
876
+ "hidden_chart_oracle_success_at_8": 0.2,
877
+ "hidden_chart_oracle_utility_at_8": 0.5298813670873642,
878
+ "ncar_to_hidden_chart_oracle_at_8": 1.4284452600977113,
879
+ "ncar_to_proposal_oracle_at_8": 0.3825324409775967,
880
+ "outcome_ptr_at_8": 0.4,
881
+ "pairwise_causal_calibration_ece": 0.24286396292452275,
882
+ "proposal_oracle_safety_label_known_at_8": 1.0,
883
+ "proposal_oracle_success_at_8": 0.0,
884
+ "proposal_oracle_success_gain_over_base_at_8": -0.2,
885
+ "proposal_oracle_unsafe_known_at_8": 0.6,
886
+ "proposal_oracle_utility_at_8": 0.11907548382878304,
887
+ "proposal_oracle_utility_gain_over_base_at_8": -0.3538366474211216,
888
+ "selected_safety_label_known_at_8": 1.0,
889
+ "selected_success_at_8": 0.0,
890
+ "selected_success_gain_over_base_at_8": -0.2,
891
+ "selected_unsafe_known_at_8": 0.4,
892
+ "selected_utility_at_8": 0.10171020515263081,
893
+ "selected_utility_gain_over_base_at_8": -0.3712019260972738,
894
+ "selector_gap_fraction_to_hidden_at_8": 0.24547848203994205,
895
+ "selector_regret_at_8": 0.01736527867615223,
896
+ "success_selector_gap_at_8": 0.0,
897
+ "success_support_gap_at_8": 0.2,
898
+ "success_total_car_to_hidden_at_8": 0.2,
899
+ "support_gap_at_8": 0.4130037985742092,
900
+ "support_gap_fraction_to_hidden_at_8": 1.2234746297648738,
901
+ "total_car_to_hidden_at_8": 0.42817116193473337
902
+ },
903
+ "2054899126": {
904
+ "any_generated_unsafe_known_at_8": 1.0,
905
+ "base_safety_label_known": 1.0,
906
+ "base_success": 1.0,
907
+ "base_unsafe_known": 0.0,
908
+ "base_utility": 2.0,
909
+ "branch_car_at_8": 0.0,
910
+ "generated_safety_label_coverage_at_8": 1.0,
911
+ "generated_unsafe_rate_known_at_8": 1.0,
912
+ "hidden_chart_oracle_success_at_8": 1.0,
913
+ "hidden_chart_oracle_utility_at_8": 2.0,
914
+ "outcome_ptr_at_8": 0.0,
915
+ "pairwise_causal_calibration_ece": 0.0,
916
+ "proposal_oracle_safety_label_known_at_8": 1.0,
917
+ "proposal_oracle_success_at_8": 1.0,
918
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
919
+ "proposal_oracle_unsafe_known_at_8": 1.0,
920
+ "proposal_oracle_utility_at_8": 2.0,
921
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
922
+ "selected_safety_label_known_at_8": 1.0,
923
+ "selected_success_at_8": 1.0,
924
+ "selected_success_gain_over_base_at_8": 0.0,
925
+ "selected_unsafe_known_at_8": 1.0,
926
+ "selected_utility_at_8": 2.0,
927
+ "selected_utility_gain_over_base_at_8": 0.0,
928
+ "selector_regret_at_8": 0.0,
929
+ "success_selector_gap_at_8": 0.0,
930
+ "success_support_gap_at_8": 0.0,
931
+ "success_total_car_to_hidden_at_8": 0.0,
932
+ "support_gap_at_8": 0.0,
933
+ "total_car_to_hidden_at_8": 0.0
934
+ },
935
+ "211245842": {
936
+ "any_generated_unsafe_known_at_8": 0.0,
937
+ "base_safety_label_known": 1.0,
938
+ "base_success": 0.0,
939
+ "base_unsafe_known": 0.0,
940
+ "base_utility": 0.791279137134552,
941
+ "branch_car_at_8": 0.04498988389968872,
942
+ "generated_safety_label_coverage_at_8": 1.0,
943
+ "generated_unsafe_rate_known_at_8": 0.0,
944
+ "hidden_chart_oracle_success_at_8": 1.0,
945
+ "hidden_chart_oracle_utility_at_8": 2.0,
946
+ "ncar_to_hidden_chart_oracle_at_8": 1.0007633933743367,
947
+ "ncar_to_proposal_oracle_at_8": 1.0209391670158559,
948
+ "outcome_ptr_at_8": 1.0,
949
+ "pairwise_causal_calibration_ece": 0.4141064151265617,
950
+ "proposal_oracle_safety_label_known_at_8": 1.0,
951
+ "proposal_oracle_success_at_8": 0.0,
952
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
953
+ "proposal_oracle_unsafe_known_at_8": 0.0,
954
+ "proposal_oracle_utility_at_8": 0.8353462815284729,
955
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0440671443939209,
956
+ "selected_safety_label_known_at_8": 1.0,
957
+ "selected_success_at_8": 0.0,
958
+ "selected_success_gain_over_base_at_8": 0.0,
959
+ "selected_unsafe_known_at_8": 0.0,
960
+ "selected_utility_at_8": 0.7903563976287842,
961
+ "selected_utility_gain_over_base_at_8": -0.0009227395057678223,
962
+ "selector_gap_fraction_to_hidden_at_8": 0.0372210700434455,
963
+ "selector_regret_at_8": 0.04498988389968872,
964
+ "success_selector_gap_at_8": 0.0,
965
+ "success_support_gap_at_8": 1.0,
966
+ "success_total_car_to_hidden_at_8": 1.0,
967
+ "support_gap_at_8": 1.164653718471527,
968
+ "support_gap_fraction_to_hidden_at_8": 0.9635423316104156,
969
+ "total_car_to_hidden_at_8": 1.2096436023712158
970
+ },
971
+ "2141802816": {
972
+ "any_generated_unsafe_known_at_8": 0.0,
973
+ "base_safety_label_known": 1.0,
974
+ "base_success": 0.0,
975
+ "base_unsafe_known": 0.0,
976
+ "base_utility": 0.23664958775043488,
977
+ "branch_car_at_8": 0.038818299770355225,
978
+ "generated_safety_label_coverage_at_8": 1.0,
979
+ "generated_unsafe_rate_known_at_8": 0.0,
980
+ "hidden_chart_oracle_success_at_8": 1.0,
981
+ "hidden_chart_oracle_utility_at_8": 2.0,
982
+ "ncar_to_hidden_chart_oracle_at_8": 1.0465578663511705,
983
+ "ncar_to_proposal_oracle_at_8": 0.8969200607232635,
984
+ "outcome_ptr_at_8": 0.0,
985
+ "pairwise_causal_calibration_ece": 0.3985289040635325,
986
+ "proposal_oracle_safety_label_known_at_8": 1.0,
987
+ "proposal_oracle_success_at_8": 0.0,
988
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
989
+ "proposal_oracle_unsafe_known_at_8": 0.0,
990
+ "proposal_oracle_utility_at_8": 0.1933700442314148,
991
+ "proposal_oracle_utility_gain_over_base_at_8": -0.04327954351902008,
992
+ "selected_safety_label_known_at_8": 1.0,
993
+ "selected_success_at_8": 0.0,
994
+ "selected_success_gain_over_base_at_8": 0.0,
995
+ "selected_unsafe_known_at_8": 0.0,
996
+ "selected_utility_at_8": 0.15455174446105957,
997
+ "selected_utility_gain_over_base_at_8": -0.0820978432893753,
998
+ "selector_gap_fraction_to_hidden_at_8": 0.022013945441980202,
999
+ "selector_regret_at_8": 0.038818299770355225,
1000
+ "success_selector_gap_at_8": 0.0,
1001
+ "success_support_gap_at_8": 1.0,
1002
+ "success_total_car_to_hidden_at_8": 1.0,
1003
+ "support_gap_at_8": 1.8066299557685852,
1004
+ "support_gap_fraction_to_hidden_at_8": 1.0245439268442436,
1005
+ "total_car_to_hidden_at_8": 1.8454482555389404
1006
+ },
1007
+ "256770874": {
1008
+ "any_generated_unsafe_known_at_8": 0.0,
1009
+ "base_safety_label_known": 1.0,
1010
+ "base_success": 0.0,
1011
+ "base_unsafe_known": 0.0,
1012
+ "base_utility": 0.18160098791122437,
1013
+ "branch_car_at_8": 0.042443424463272095,
1014
+ "generated_safety_label_coverage_at_8": 1.0,
1015
+ "generated_unsafe_rate_known_at_8": 0.0,
1016
+ "hidden_chart_oracle_success_at_8": 1.0,
1017
+ "hidden_chart_oracle_utility_at_8": 2.0,
1018
+ "ncar_to_hidden_chart_oracle_at_8": 1.002821760012782,
1019
+ "ncar_to_proposal_oracle_at_8": 1.1375171273285567,
1020
+ "outcome_ptr_at_8": 1.0,
1021
+ "pairwise_causal_calibration_ece": 0.26681470238264204,
1022
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1023
+ "proposal_oracle_success_at_8": 0.0,
1024
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1025
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1026
+ "proposal_oracle_utility_at_8": 0.21891331672668457,
1027
+ "proposal_oracle_utility_gain_over_base_at_8": 0.037312328815460205,
1028
+ "selected_safety_label_known_at_8": 1.0,
1029
+ "selected_success_at_8": 0.0,
1030
+ "selected_success_gain_over_base_at_8": 0.0,
1031
+ "selected_unsafe_known_at_8": 0.0,
1032
+ "selected_utility_at_8": 0.17646989226341248,
1033
+ "selected_utility_gain_over_base_at_8": -0.00513109564781189,
1034
+ "selector_gap_fraction_to_hidden_at_8": 0.023341095205786425,
1035
+ "selector_regret_at_8": 0.042443424463272095,
1036
+ "success_selector_gap_at_8": 0.0,
1037
+ "success_support_gap_at_8": 1.0,
1038
+ "success_total_car_to_hidden_at_8": 1.0,
1039
+ "support_gap_at_8": 1.7810866832733154,
1040
+ "support_gap_fraction_to_hidden_at_8": 0.9794806703218564,
1041
+ "total_car_to_hidden_at_8": 1.8235301077365875
1042
+ },
1043
+ "278008524": {
1044
+ "any_generated_unsafe_known_at_8": 0.5,
1045
+ "base_safety_label_known": 1.0,
1046
+ "base_success": 0.0,
1047
+ "base_unsafe_known": 0.0,
1048
+ "base_utility": 0.060655687004327774,
1049
+ "branch_car_at_8": 0.008101326413452625,
1050
+ "generated_safety_label_coverage_at_8": 1.0,
1051
+ "generated_unsafe_rate_known_at_8": 0.5,
1052
+ "hidden_chart_oracle_success_at_8": 0.5,
1053
+ "hidden_chart_oracle_utility_at_8": 1.113811194896698,
1054
+ "ncar_to_hidden_chart_oracle_at_8": 1.1481789711765635,
1055
+ "ncar_to_proposal_oracle_at_8": 0.4289430850076562,
1056
+ "outcome_ptr_at_8": 0.0,
1057
+ "pairwise_causal_calibration_ece": 0.20536892413092878,
1058
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1059
+ "proposal_oracle_success_at_8": 0.0,
1060
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1061
+ "proposal_oracle_unsafe_known_at_8": 0.5,
1062
+ "proposal_oracle_utility_at_8": 0.03437099326401949,
1063
+ "proposal_oracle_utility_gain_over_base_at_8": -0.026284693740308285,
1064
+ "selected_safety_label_known_at_8": 1.0,
1065
+ "selected_success_at_8": 0.0,
1066
+ "selected_success_gain_over_base_at_8": 0.0,
1067
+ "selected_unsafe_known_at_8": 0.5,
1068
+ "selected_utility_at_8": 0.026269666850566864,
1069
+ "selected_utility_gain_over_base_at_8": -0.03438602015376091,
1070
+ "selector_gap_fraction_to_hidden_at_8": 0.024682996062310017,
1071
+ "selector_regret_at_8": 0.008101326413452625,
1072
+ "success_selector_gap_at_8": 0.0,
1073
+ "success_support_gap_at_8": 0.5,
1074
+ "success_total_car_to_hidden_at_8": 0.5,
1075
+ "support_gap_at_8": 1.0794402016326785,
1076
+ "support_gap_fraction_to_hidden_at_8": 1.1234960169117114,
1077
+ "total_car_to_hidden_at_8": 1.0875415280461311
1078
+ },
1079
+ "416183106": {
1080
+ "any_generated_unsafe_known_at_8": 0.0,
1081
+ "base_safety_label_known": 1.0,
1082
+ "base_success": 0.0,
1083
+ "base_unsafe_known": 0.0,
1084
+ "base_utility": 0.23001998662948608,
1085
+ "branch_car_at_8": 0.0,
1086
+ "generated_safety_label_coverage_at_8": 1.0,
1087
+ "generated_unsafe_rate_known_at_8": 0.0,
1088
+ "hidden_chart_oracle_success_at_8": 1.0,
1089
+ "hidden_chart_oracle_utility_at_8": 2.0,
1090
+ "ncar_to_hidden_chart_oracle_at_8": 0.828016730744017,
1091
+ "ncar_to_proposal_oracle_at_8": 0.0,
1092
+ "outcome_ptr_at_8": 1.0,
1093
+ "pairwise_causal_calibration_ece": 0.07874347996488716,
1094
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1095
+ "proposal_oracle_success_at_8": 0.0,
1096
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1097
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1098
+ "proposal_oracle_utility_at_8": 0.5344269275665283,
1099
+ "proposal_oracle_utility_gain_over_base_at_8": 0.30440694093704224,
1100
+ "selected_safety_label_known_at_8": 1.0,
1101
+ "selected_success_at_8": 0.0,
1102
+ "selected_success_gain_over_base_at_8": 0.0,
1103
+ "selected_unsafe_known_at_8": 0.0,
1104
+ "selected_utility_at_8": 0.5344269275665283,
1105
+ "selected_utility_gain_over_base_at_8": 0.30440694093704224,
1106
+ "selector_gap_fraction_to_hidden_at_8": 0.0,
1107
+ "selector_regret_at_8": 0.0,
1108
+ "success_selector_gap_at_8": 0.0,
1109
+ "success_support_gap_at_8": 1.0,
1110
+ "success_total_car_to_hidden_at_8": 1.0,
1111
+ "support_gap_at_8": 1.4655730724334717,
1112
+ "support_gap_fraction_to_hidden_at_8": 0.8280167354221304,
1113
+ "total_car_to_hidden_at_8": 1.4655730724334717
1114
+ },
1115
+ "534717141": {
1116
+ "any_generated_unsafe_known_at_8": 0.0,
1117
+ "base_safety_label_known": 1.0,
1118
+ "base_success": 0.0,
1119
+ "base_unsafe_known": 0.0,
1120
+ "base_utility": 0.13276933133602142,
1121
+ "branch_car_at_8": 0.0,
1122
+ "generated_safety_label_coverage_at_8": 1.0,
1123
+ "generated_unsafe_rate_known_at_8": 0.0,
1124
+ "hidden_chart_oracle_success_at_8": 1.0,
1125
+ "hidden_chart_oracle_utility_at_8": 2.0,
1126
+ "ncar_to_hidden_chart_oracle_at_8": 1.0230647928623209,
1127
+ "ncar_to_proposal_oracle_at_8": 0.0,
1128
+ "outcome_ptr_at_8": 0.0,
1129
+ "pairwise_causal_calibration_ece": 0.0753952117725891,
1130
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1131
+ "proposal_oracle_success_at_8": 0.0,
1132
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1133
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1134
+ "proposal_oracle_utility_at_8": 0.08970203250646591,
1135
+ "proposal_oracle_utility_gain_over_base_at_8": -0.04306729882955551,
1136
+ "selected_safety_label_known_at_8": 1.0,
1137
+ "selected_success_at_8": 0.0,
1138
+ "selected_success_gain_over_base_at_8": 0.0,
1139
+ "selected_unsafe_known_at_8": 0.0,
1140
+ "selected_utility_at_8": 0.08970203250646591,
1141
+ "selected_utility_gain_over_base_at_8": -0.04306729882955551,
1142
+ "selector_gap_fraction_to_hidden_at_8": 0.0,
1143
+ "selector_regret_at_8": 0.0,
1144
+ "success_selector_gap_at_8": 0.0,
1145
+ "success_support_gap_at_8": 1.0,
1146
+ "success_total_car_to_hidden_at_8": 1.0,
1147
+ "support_gap_at_8": 1.910297967493534,
1148
+ "support_gap_fraction_to_hidden_at_8": 1.0230647983413697,
1149
+ "total_car_to_hidden_at_8": 1.910297967493534
1150
+ },
1151
+ "551504635": {
1152
+ "any_generated_unsafe_known_at_8": 0.0,
1153
+ "base_safety_label_known": 1.0,
1154
+ "base_success": 0.0,
1155
+ "base_unsafe_known": 0.0,
1156
+ "base_utility": 0.7272467017173767,
1157
+ "branch_car_at_8": 1.2730597853660583,
1158
+ "generated_safety_label_coverage_at_8": 1.0,
1159
+ "generated_unsafe_rate_known_at_8": 0.0,
1160
+ "hidden_chart_oracle_success_at_8": 1.0,
1161
+ "hidden_chart_oracle_utility_at_8": 2.0,
1162
+ "ncar_to_hidden_chart_oracle_at_8": 1.0002407984968027,
1163
+ "ncar_to_proposal_oracle_at_8": 1.0002407984968027,
1164
+ "outcome_ptr_at_8": 1.0,
1165
+ "pairwise_causal_calibration_ece": 0.2344395807565722,
1166
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1167
+ "proposal_oracle_success_at_8": 1.0,
1168
+ "proposal_oracle_success_gain_over_base_at_8": 1.0,
1169
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1170
+ "proposal_oracle_utility_at_8": 2.0,
1171
+ "proposal_oracle_utility_gain_over_base_at_8": 1.2727532982826233,
1172
+ "selected_safety_label_known_at_8": 1.0,
1173
+ "selected_success_at_8": 0.0,
1174
+ "selected_success_gain_over_base_at_8": 0.0,
1175
+ "selected_unsafe_known_at_8": 0.0,
1176
+ "selected_utility_at_8": 0.7269402146339417,
1177
+ "selected_utility_gain_over_base_at_8": -0.0003064870834350586,
1178
+ "selector_gap_fraction_to_hidden_at_8": 1.0002408063556767,
1179
+ "selector_regret_at_8": 1.2730597853660583,
1180
+ "success_selector_gap_at_8": 1.0,
1181
+ "success_support_gap_at_8": 0.0,
1182
+ "success_total_car_to_hidden_at_8": 1.0,
1183
+ "support_gap_at_8": 0.0,
1184
+ "support_gap_fraction_to_hidden_at_8": 0.0,
1185
+ "total_car_to_hidden_at_8": 1.2730597853660583
1186
+ },
1187
+ "565369392": {
1188
+ "any_generated_unsafe_known_at_8": 1.0,
1189
+ "base_safety_label_known": 1.0,
1190
+ "base_success": 0.3333333333333333,
1191
+ "base_unsafe_known": 0.0,
1192
+ "base_utility": 0.8445292512575785,
1193
+ "branch_car_at_8": 0.0,
1194
+ "generated_safety_label_coverage_at_8": 1.0,
1195
+ "generated_unsafe_rate_known_at_8": 0.8333333333333334,
1196
+ "hidden_chart_oracle_success_at_8": 0.6666666666666666,
1197
+ "hidden_chart_oracle_utility_at_8": 1.4266315797964733,
1198
+ "ncar_to_hidden_chart_oracle_at_8": 4.740664659244946,
1199
+ "ncar_to_proposal_oracle_at_8": 0.0,
1200
+ "outcome_ptr_at_8": 0.3333333333333333,
1201
+ "pairwise_causal_calibration_ece": 0.13140272637925274,
1202
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1203
+ "proposal_oracle_success_at_8": 0.0,
1204
+ "proposal_oracle_success_gain_over_base_at_8": -0.3333333333333333,
1205
+ "proposal_oracle_unsafe_known_at_8": 0.6666666666666666,
1206
+ "proposal_oracle_utility_at_8": 0.20857956757148108,
1207
+ "proposal_oracle_utility_gain_over_base_at_8": -0.6359496836860975,
1208
+ "selected_safety_label_known_at_8": 1.0,
1209
+ "selected_success_at_8": 0.0,
1210
+ "selected_success_gain_over_base_at_8": -0.3333333333333333,
1211
+ "selected_unsafe_known_at_8": 0.6666666666666666,
1212
+ "selected_utility_at_8": 0.20857956757148108,
1213
+ "selected_utility_gain_over_base_at_8": -0.6359496836860975,
1214
+ "selector_gap_fraction_to_hidden_at_8": 0.0,
1215
+ "selector_regret_at_8": 0.0,
1216
+ "success_selector_gap_at_8": 0.0,
1217
+ "success_support_gap_at_8": 0.6666666666666666,
1218
+ "success_total_car_to_hidden_at_8": 0.6666666666666666,
1219
+ "support_gap_at_8": 1.218052012224992,
1220
+ "support_gap_fraction_to_hidden_at_8": 4.740673332898252,
1221
+ "total_car_to_hidden_at_8": 1.218052012224992
1222
+ },
1223
+ "565621826": {
1224
+ "any_generated_unsafe_known_at_8": 1.0,
1225
+ "base_safety_label_known": 1.0,
1226
+ "base_success": 1.0,
1227
+ "base_unsafe_known": 0.0,
1228
+ "base_utility": 2.0,
1229
+ "branch_car_at_8": 0.0,
1230
+ "generated_safety_label_coverage_at_8": 1.0,
1231
+ "generated_unsafe_rate_known_at_8": 1.0,
1232
+ "hidden_chart_oracle_success_at_8": 1.0,
1233
+ "hidden_chart_oracle_utility_at_8": 2.0,
1234
+ "outcome_ptr_at_8": 0.0,
1235
+ "pairwise_causal_calibration_ece": 0.0,
1236
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1237
+ "proposal_oracle_success_at_8": 1.0,
1238
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1239
+ "proposal_oracle_unsafe_known_at_8": 1.0,
1240
+ "proposal_oracle_utility_at_8": 2.0,
1241
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
1242
+ "selected_safety_label_known_at_8": 1.0,
1243
+ "selected_success_at_8": 1.0,
1244
+ "selected_success_gain_over_base_at_8": 0.0,
1245
+ "selected_unsafe_known_at_8": 1.0,
1246
+ "selected_utility_at_8": 2.0,
1247
+ "selected_utility_gain_over_base_at_8": 0.0,
1248
+ "selector_regret_at_8": 0.0,
1249
+ "success_selector_gap_at_8": 0.0,
1250
+ "success_support_gap_at_8": 0.0,
1251
+ "success_total_car_to_hidden_at_8": 0.0,
1252
+ "support_gap_at_8": 0.0,
1253
+ "total_car_to_hidden_at_8": 0.0
1254
+ },
1255
+ "585904130": {
1256
+ "any_generated_unsafe_known_at_8": 0.3333333333333333,
1257
+ "base_safety_label_known": 1.0,
1258
+ "base_success": 0.0,
1259
+ "base_unsafe_known": 0.0,
1260
+ "base_utility": 0.13642114897569022,
1261
+ "branch_car_at_8": 0.022781788061062496,
1262
+ "generated_safety_label_coverage_at_8": 1.0,
1263
+ "generated_unsafe_rate_known_at_8": 0.3333333333333333,
1264
+ "hidden_chart_oracle_success_at_8": 0.0,
1265
+ "hidden_chart_oracle_utility_at_8": 0.23764699200789133,
1266
+ "ncar_to_hidden_chart_oracle_at_8": 1.5916216907065317,
1267
+ "ncar_to_proposal_oracle_at_8": 0.5193742667883882,
1268
+ "outcome_ptr_at_8": 0.3333333333333333,
1269
+ "pairwise_causal_calibration_ece": 0.17283993917588428,
1270
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1271
+ "proposal_oracle_success_at_8": 0.0,
1272
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1273
+ "proposal_oracle_unsafe_known_at_8": 0.3333333333333333,
1274
+ "proposal_oracle_utility_at_8": 0.09142490103840828,
1275
+ "proposal_oracle_utility_gain_over_base_at_8": -0.04499624793728193,
1276
+ "selected_safety_label_known_at_8": 1.0,
1277
+ "selected_success_at_8": 0.0,
1278
+ "selected_success_gain_over_base_at_8": 0.0,
1279
+ "selected_unsafe_known_at_8": 0.3333333333333333,
1280
+ "selected_utility_at_8": 0.06864311297734578,
1281
+ "selected_utility_gain_over_base_at_8": -0.06777803599834442,
1282
+ "selector_gap_fraction_to_hidden_at_8": 0.37677313988714195,
1283
+ "selector_regret_at_8": 0.022781788061062496,
1284
+ "success_selector_gap_at_8": 0.0,
1285
+ "success_support_gap_at_8": 0.0,
1286
+ "success_total_car_to_hidden_at_8": 0.0,
1287
+ "support_gap_at_8": 0.14622209096948305,
1288
+ "support_gap_fraction_to_hidden_at_8": 1.2148487255248963,
1289
+ "total_car_to_hidden_at_8": 0.16900387903054556
1290
+ },
1291
+ "626486813": {
1292
+ "any_generated_unsafe_known_at_8": 0.5,
1293
+ "base_safety_label_known": 1.0,
1294
+ "base_success": 0.25,
1295
+ "base_unsafe_known": 0.0,
1296
+ "base_utility": 0.5452363695949316,
1297
+ "branch_car_at_8": 0.01009321678429842,
1298
+ "generated_safety_label_coverage_at_8": 1.0,
1299
+ "generated_unsafe_rate_known_at_8": 0.28125,
1300
+ "hidden_chart_oracle_success_at_8": 0.25,
1301
+ "hidden_chart_oracle_utility_at_8": 0.6124684102833271,
1302
+ "ncar_to_hidden_chart_oracle_at_8": 0.8102411086604181,
1303
+ "ncar_to_proposal_oracle_at_8": 0.2258722711825077,
1304
+ "outcome_ptr_at_8": 0.25,
1305
+ "pairwise_causal_calibration_ece": 0.2769376371795747,
1306
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1307
+ "proposal_oracle_success_at_8": 0.0,
1308
+ "proposal_oracle_success_gain_over_base_at_8": -0.25,
1309
+ "proposal_oracle_unsafe_known_at_8": 0.25,
1310
+ "proposal_oracle_utility_at_8": 0.09146204218268394,
1311
+ "proposal_oracle_utility_gain_over_base_at_8": -0.45377432741224766,
1312
+ "selected_safety_label_known_at_8": 1.0,
1313
+ "selected_success_at_8": 0.0,
1314
+ "selected_success_gain_over_base_at_8": -0.25,
1315
+ "selected_unsafe_known_at_8": 0.25,
1316
+ "selected_utility_at_8": 0.08136882539838552,
1317
+ "selected_utility_gain_over_base_at_8": -0.4638675441965461,
1318
+ "selector_gap_fraction_to_hidden_at_8": 0.3592836940693531,
1319
+ "selector_regret_at_8": 0.01009321678429842,
1320
+ "success_selector_gap_at_8": 0.0,
1321
+ "success_support_gap_at_8": 0.25,
1322
+ "success_total_car_to_hidden_at_8": 0.25,
1323
+ "support_gap_at_8": 0.5353103466331959,
1324
+ "support_gap_fraction_to_hidden_at_8": 0.7858811145047201,
1325
+ "total_car_to_hidden_at_8": 0.5373998517170548
1326
+ },
1327
+ "636680137": {
1328
+ "any_generated_unsafe_known_at_8": 1.0,
1329
+ "base_safety_label_known": 1.0,
1330
+ "base_success": 0.0,
1331
+ "base_unsafe_known": 0.0,
1332
+ "base_utility": 0.07024133205413818,
1333
+ "branch_car_at_8": 0.02178390882909298,
1334
+ "generated_safety_label_coverage_at_8": 1.0,
1335
+ "generated_unsafe_rate_known_at_8": 0.25,
1336
+ "hidden_chart_oracle_success_at_8": 1.0,
1337
+ "hidden_chart_oracle_utility_at_8": 2.0,
1338
+ "ncar_to_hidden_chart_oracle_at_8": 1.0226806271202884,
1339
+ "ncar_to_proposal_oracle_at_8": 0.9908871412464478,
1340
+ "outcome_ptr_at_8": 0.0,
1341
+ "pairwise_causal_calibration_ece": 0.5932483433892377,
1342
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1343
+ "proposal_oracle_success_at_8": 0.0,
1344
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1345
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1346
+ "proposal_oracle_utility_at_8": 0.048257093876600266,
1347
+ "proposal_oracle_utility_gain_over_base_at_8": -0.021984238177537918,
1348
+ "selected_safety_label_known_at_8": 1.0,
1349
+ "selected_success_at_8": 0.0,
1350
+ "selected_success_gain_over_base_at_8": 0.0,
1351
+ "selected_unsafe_known_at_8": 0.0,
1352
+ "selected_utility_at_8": 0.026473185047507286,
1353
+ "selected_utility_gain_over_base_at_8": -0.0437681470066309,
1354
+ "selector_gap_fraction_to_hidden_at_8": 0.011288410924605892,
1355
+ "selector_regret_at_8": 0.02178390882909298,
1356
+ "success_selector_gap_at_8": 0.0,
1357
+ "success_support_gap_at_8": 1.0,
1358
+ "success_total_car_to_hidden_at_8": 1.0,
1359
+ "support_gap_at_8": 1.9517429061233997,
1360
+ "support_gap_fraction_to_hidden_at_8": 1.0113922214952087,
1361
+ "total_car_to_hidden_at_8": 1.9735268149524927
1362
+ },
1363
+ "706479600": {
1364
+ "any_generated_unsafe_known_at_8": 0.0,
1365
+ "base_safety_label_known": 1.0,
1366
+ "base_success": 0.0,
1367
+ "base_unsafe_known": 0.0,
1368
+ "base_utility": 0.8188905119895935,
1369
+ "branch_car_at_8": 0.018285930156707764,
1370
+ "generated_safety_label_coverage_at_8": 1.0,
1371
+ "generated_unsafe_rate_known_at_8": 0.0,
1372
+ "hidden_chart_oracle_success_at_8": 1.0,
1373
+ "hidden_chart_oracle_utility_at_8": 2.0,
1374
+ "ncar_to_hidden_chart_oracle_at_8": 1.0001120742118061,
1375
+ "ncar_to_proposal_oracle_at_8": 1.0072917889865605,
1376
+ "outcome_ptr_at_8": 1.0,
1377
+ "pairwise_causal_calibration_ece": 0.28238485246274814,
1378
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1379
+ "proposal_oracle_success_at_8": 0.0,
1380
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1381
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1382
+ "proposal_oracle_utility_at_8": 0.8370440602302551,
1383
+ "proposal_oracle_utility_gain_over_base_at_8": 0.01815354824066162,
1384
+ "selected_safety_label_known_at_8": 1.0,
1385
+ "selected_success_at_8": 0.0,
1386
+ "selected_success_gain_over_base_at_8": 0.0,
1387
+ "selected_unsafe_known_at_8": 0.0,
1388
+ "selected_utility_at_8": 0.8187581300735474,
1389
+ "selected_utility_gain_over_base_at_8": -0.00013238191604614258,
1390
+ "selector_gap_fraction_to_hidden_at_8": 0.015481994126988717,
1391
+ "selector_regret_at_8": 0.018285930156707764,
1392
+ "success_selector_gap_at_8": 0.0,
1393
+ "success_support_gap_at_8": 1.0,
1394
+ "success_total_car_to_hidden_at_8": 1.0,
1395
+ "support_gap_at_8": 1.1629559397697449,
1396
+ "support_gap_fraction_to_hidden_at_8": 0.9846300885523818,
1397
+ "total_car_to_hidden_at_8": 1.1812418699264526
1398
+ },
1399
+ "750448296": {
1400
+ "any_generated_unsafe_known_at_8": 0.5,
1401
+ "base_safety_label_known": 1.0,
1402
+ "base_success": 0.0,
1403
+ "base_unsafe_known": 0.0,
1404
+ "base_utility": 0.14706942811608315,
1405
+ "branch_car_at_8": 0.02600152976810932,
1406
+ "generated_safety_label_coverage_at_8": 1.0,
1407
+ "generated_unsafe_rate_known_at_8": 0.1875,
1408
+ "hidden_chart_oracle_success_at_8": 0.0,
1409
+ "hidden_chart_oracle_utility_at_8": 0.23580759018659592,
1410
+ "ncar_to_hidden_chart_oracle_at_8": 4.978844987783829,
1411
+ "ncar_to_proposal_oracle_at_8": 5.576871543359603,
1412
+ "outcome_ptr_at_8": 0.5,
1413
+ "pairwise_causal_calibration_ece": 0.19889074905685378,
1414
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1415
+ "proposal_oracle_success_at_8": 0.0,
1416
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1417
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1418
+ "proposal_oracle_utility_at_8": 0.14822042733430862,
1419
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0011509992182254791,
1420
+ "selected_safety_label_known_at_8": 1.0,
1421
+ "selected_success_at_8": 0.0,
1422
+ "selected_success_gain_over_base_at_8": 0.0,
1423
+ "selected_unsafe_known_at_8": 0.0,
1424
+ "selected_utility_at_8": 0.1222188975661993,
1425
+ "selected_utility_gain_over_base_at_8": -0.024850530549883842,
1426
+ "selector_gap_fraction_to_hidden_at_8": 3.5662633342346166,
1427
+ "selector_regret_at_8": 0.02600152976810932,
1428
+ "success_selector_gap_at_8": 0.0,
1429
+ "success_support_gap_at_8": 0.0,
1430
+ "success_total_car_to_hidden_at_8": 0.0,
1431
+ "support_gap_at_8": 0.08758716285228729,
1432
+ "support_gap_fraction_to_hidden_at_8": 1.41259115683824,
1433
+ "total_car_to_hidden_at_8": 0.11358869262039661
1434
+ },
1435
+ "75070641": {
1436
+ "any_generated_unsafe_known_at_8": 0.0,
1437
+ "base_safety_label_known": 1.0,
1438
+ "base_success": 0.0,
1439
+ "base_unsafe_known": 0.0,
1440
+ "base_utility": 0.194026380777359,
1441
+ "branch_car_at_8": 0.024178534746170044,
1442
+ "generated_safety_label_coverage_at_8": 1.0,
1443
+ "generated_unsafe_rate_known_at_8": 0.0,
1444
+ "hidden_chart_oracle_success_at_8": 1.0,
1445
+ "hidden_chart_oracle_utility_at_8": 2.0,
1446
+ "ncar_to_hidden_chart_oracle_at_8": 1.0008596290963836,
1447
+ "ncar_to_proposal_oracle_at_8": 1.0686141104703737,
1448
+ "outcome_ptr_at_8": 1.0,
1449
+ "pairwise_causal_calibration_ece": 0.1631929231601434,
1450
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1451
+ "proposal_oracle_success_at_8": 0.0,
1452
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1453
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1454
+ "proposal_oracle_utility_at_8": 0.21665243804454803,
1455
+ "proposal_oracle_utility_gain_over_base_at_8": 0.022626057267189026,
1456
+ "selected_safety_label_known_at_8": 1.0,
1457
+ "selected_success_at_8": 0.0,
1458
+ "selected_success_gain_over_base_at_8": 0.0,
1459
+ "selected_unsafe_known_at_8": 0.0,
1460
+ "selected_utility_at_8": 0.192473903298378,
1461
+ "selected_utility_gain_over_base_at_8": -0.001552477478981018,
1462
+ "selector_gap_fraction_to_hidden_at_8": 0.01338808855722787,
1463
+ "selector_regret_at_8": 0.024178534746170044,
1464
+ "success_selector_gap_at_8": 0.0,
1465
+ "success_support_gap_at_8": 1.0,
1466
+ "success_total_car_to_hidden_at_8": 1.0,
1467
+ "support_gap_at_8": 1.783347561955452,
1468
+ "support_gap_fraction_to_hidden_at_8": 0.987471546081095,
1469
+ "total_car_to_hidden_at_8": 1.807526096701622
1470
+ },
1471
+ "767747424": {
1472
+ "any_generated_unsafe_known_at_8": 0.5,
1473
+ "base_safety_label_known": 1.0,
1474
+ "base_success": 0.0,
1475
+ "base_unsafe_known": 0.0,
1476
+ "base_utility": 0.10894576646387577,
1477
+ "branch_car_at_8": 0.016496962402015924,
1478
+ "generated_safety_label_coverage_at_8": 1.0,
1479
+ "generated_unsafe_rate_known_at_8": 0.5,
1480
+ "hidden_chart_oracle_success_at_8": 0.0,
1481
+ "hidden_chart_oracle_utility_at_8": 0.18863565661013126,
1482
+ "ncar_to_hidden_chart_oracle_at_8": 5.205313661372345,
1483
+ "ncar_to_proposal_oracle_at_8": 0.9043237597220475,
1484
+ "outcome_ptr_at_8": 0.5,
1485
+ "pairwise_causal_calibration_ece": 0.24962611487178488,
1486
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1487
+ "proposal_oracle_success_at_8": 0.0,
1488
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1489
+ "proposal_oracle_unsafe_known_at_8": 0.5,
1490
+ "proposal_oracle_utility_at_8": 0.10893216636031866,
1491
+ "proposal_oracle_utility_gain_over_base_at_8": -1.3600103557109833e-05,
1492
+ "selected_safety_label_known_at_8": 1.0,
1493
+ "selected_success_at_8": 0.0,
1494
+ "selected_success_gain_over_base_at_8": 0.0,
1495
+ "selected_unsafe_known_at_8": 0.5,
1496
+ "selected_utility_at_8": 0.09243520395830274,
1497
+ "selected_utility_gain_over_base_at_8": -0.016510562505573034,
1498
+ "selector_gap_fraction_to_hidden_at_8": 2.999421934457336,
1499
+ "selector_regret_at_8": 0.016496962402015924,
1500
+ "success_selector_gap_at_8": 0.0,
1501
+ "success_support_gap_at_8": 0.0,
1502
+ "success_total_car_to_hidden_at_8": 0.0,
1503
+ "support_gap_at_8": 0.0797034902498126,
1504
+ "support_gap_fraction_to_hidden_at_8": 2.2059097841347444,
1505
+ "total_car_to_hidden_at_8": 0.09620045265182853
1506
+ },
1507
+ "77147264": {
1508
+ "any_generated_unsafe_known_at_8": 0.0,
1509
+ "base_safety_label_known": 1.0,
1510
+ "base_success": 0.5,
1511
+ "base_unsafe_known": 0.0,
1512
+ "base_utility": 1.0903030335903168,
1513
+ "branch_car_at_8": 0.00890406221151352,
1514
+ "generated_safety_label_coverage_at_8": 1.0,
1515
+ "generated_unsafe_rate_known_at_8": 0.0,
1516
+ "hidden_chart_oracle_success_at_8": 1.0,
1517
+ "hidden_chart_oracle_utility_at_8": 2.0,
1518
+ "ncar_to_hidden_chart_oracle_at_8": 1.0010257406519596,
1519
+ "ncar_to_proposal_oracle_at_8": 1.1170642490674998,
1520
+ "outcome_ptr_at_8": 0.5,
1521
+ "pairwise_causal_calibration_ece": 0.05996298782788377,
1522
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1523
+ "proposal_oracle_success_at_8": 0.5,
1524
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1525
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1526
+ "proposal_oracle_utility_at_8": 1.098273977637291,
1527
+ "proposal_oracle_utility_gain_over_base_at_8": 0.007970944046974182,
1528
+ "selected_safety_label_known_at_8": 1.0,
1529
+ "selected_success_at_8": 0.5,
1530
+ "selected_success_gain_over_base_at_8": 0.0,
1531
+ "selected_unsafe_known_at_8": 0.0,
1532
+ "selected_utility_at_8": 1.0893699154257774,
1533
+ "selected_utility_gain_over_base_at_8": -0.0009331181645393372,
1534
+ "selector_gap_fraction_to_hidden_at_8": 0.009787943172609816,
1535
+ "selector_regret_at_8": 0.00890406221151352,
1536
+ "success_selector_gap_at_8": 0.0,
1537
+ "success_support_gap_at_8": 0.5,
1538
+ "success_total_car_to_hidden_at_8": 0.5,
1539
+ "support_gap_at_8": 0.901726022362709,
1540
+ "support_gap_fraction_to_hidden_at_8": 0.9912378029813232,
1541
+ "total_car_to_hidden_at_8": 0.9106300845742226
1542
+ },
1543
+ "794181626": {
1544
+ "any_generated_unsafe_known_at_8": 0.0,
1545
+ "base_safety_label_known": 1.0,
1546
+ "base_success": 1.0,
1547
+ "base_unsafe_known": 0.0,
1548
+ "base_utility": 2.0,
1549
+ "branch_car_at_8": 0.0,
1550
+ "generated_safety_label_coverage_at_8": 1.0,
1551
+ "generated_unsafe_rate_known_at_8": 0.0,
1552
+ "hidden_chart_oracle_success_at_8": 1.0,
1553
+ "hidden_chart_oracle_utility_at_8": 2.0,
1554
+ "outcome_ptr_at_8": 0.0,
1555
+ "pairwise_causal_calibration_ece": 0.0,
1556
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1557
+ "proposal_oracle_success_at_8": 1.0,
1558
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1559
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1560
+ "proposal_oracle_utility_at_8": 2.0,
1561
+ "proposal_oracle_utility_gain_over_base_at_8": 0.0,
1562
+ "selected_safety_label_known_at_8": 1.0,
1563
+ "selected_success_at_8": 1.0,
1564
+ "selected_success_gain_over_base_at_8": 0.0,
1565
+ "selected_unsafe_known_at_8": 0.0,
1566
+ "selected_utility_at_8": 2.0,
1567
+ "selected_utility_gain_over_base_at_8": 0.0,
1568
+ "selector_regret_at_8": 0.0,
1569
+ "success_selector_gap_at_8": 0.0,
1570
+ "success_support_gap_at_8": 0.0,
1571
+ "success_total_car_to_hidden_at_8": 0.0,
1572
+ "support_gap_at_8": 0.0,
1573
+ "total_car_to_hidden_at_8": 0.0
1574
+ },
1575
+ "957101489": {
1576
+ "any_generated_unsafe_known_at_8": 0.0,
1577
+ "base_safety_label_known": 1.0,
1578
+ "base_success": 0.0,
1579
+ "base_unsafe_known": 0.0,
1580
+ "base_utility": 0.23217551410198212,
1581
+ "branch_car_at_8": 0.3351810723543167,
1582
+ "generated_safety_label_coverage_at_8": 1.0,
1583
+ "generated_unsafe_rate_known_at_8": 0.0,
1584
+ "hidden_chart_oracle_success_at_8": 1.0,
1585
+ "hidden_chart_oracle_utility_at_8": 2.0,
1586
+ "ncar_to_hidden_chart_oracle_at_8": 0.998510000020915,
1587
+ "ncar_to_proposal_oracle_at_8": 0.9922026628719985,
1588
+ "outcome_ptr_at_8": 1.0,
1589
+ "pairwise_causal_calibration_ece": 0.4692653473226291,
1590
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1591
+ "proposal_oracle_success_at_8": 0.0,
1592
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1593
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1594
+ "proposal_oracle_utility_at_8": 0.5699906349182129,
1595
+ "proposal_oracle_utility_gain_over_base_at_8": 0.3378151208162308,
1596
+ "selected_safety_label_known_at_8": 1.0,
1597
+ "selected_success_at_8": 0.0,
1598
+ "selected_success_gain_over_base_at_8": 0.0,
1599
+ "selected_unsafe_known_at_8": 0.0,
1600
+ "selected_utility_at_8": 0.23480956256389618,
1601
+ "selected_utility_gain_over_base_at_8": 0.0026340484619140625,
1602
+ "selector_gap_fraction_to_hidden_at_8": 0.18960087668660824,
1603
+ "selector_regret_at_8": 0.3351810723543167,
1604
+ "success_selector_gap_at_8": 0.0,
1605
+ "success_support_gap_at_8": 1.0,
1606
+ "success_total_car_to_hidden_at_8": 1.0,
1607
+ "support_gap_at_8": 1.430009365081787,
1608
+ "support_gap_fraction_to_hidden_at_8": 0.8089091289825484,
1609
+ "total_car_to_hidden_at_8": 1.7651904374361038
1610
+ },
1611
+ "998761162": {
1612
+ "any_generated_unsafe_known_at_8": 0.0,
1613
+ "base_safety_label_known": 1.0,
1614
+ "base_success": 0.0,
1615
+ "base_unsafe_known": 0.0,
1616
+ "base_utility": 0.20425210893154144,
1617
+ "branch_car_at_8": 0.004930347204208374,
1618
+ "generated_safety_label_coverage_at_8": 1.0,
1619
+ "generated_unsafe_rate_known_at_8": 0.0,
1620
+ "hidden_chart_oracle_success_at_8": 1.0,
1621
+ "hidden_chart_oracle_utility_at_8": 2.0,
1622
+ "ncar_to_hidden_chart_oracle_at_8": 1.062306753304616,
1623
+ "ncar_to_proposal_oracle_at_8": 0.04609658160039006,
1624
+ "outcome_ptr_at_8": 0.0,
1625
+ "pairwise_causal_calibration_ece": 0.24317634255647042,
1626
+ "proposal_oracle_safety_label_known_at_8": 1.0,
1627
+ "proposal_oracle_success_at_8": 0.0,
1628
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
1629
+ "proposal_oracle_unsafe_known_at_8": 0.0,
1630
+ "proposal_oracle_utility_at_8": 0.09729522466659546,
1631
+ "proposal_oracle_utility_gain_over_base_at_8": -0.10695688426494598,
1632
+ "selected_safety_label_known_at_8": 1.0,
1633
+ "selected_success_at_8": 0.0,
1634
+ "selected_success_gain_over_base_at_8": 0.0,
1635
+ "selected_unsafe_known_at_8": 0.0,
1636
+ "selected_utility_at_8": 0.09236487746238708,
1637
+ "selected_utility_gain_over_base_at_8": -0.11188723146915436,
1638
+ "selector_gap_fraction_to_hidden_at_8": 0.002745567587037426,
1639
+ "selector_regret_at_8": 0.004930347204208374,
1640
+ "success_selector_gap_at_8": 0.0,
1641
+ "success_support_gap_at_8": 1.0,
1642
+ "success_total_car_to_hidden_at_8": 1.0,
1643
+ "support_gap_at_8": 1.9027047753334045,
1644
+ "support_gap_fraction_to_hidden_at_8": 1.0595611916332572,
1645
+ "total_car_to_hidden_at_8": 1.907635122537613
1646
+ }
1647
+ }
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/metrics_by_task.json ADDED
@@ -0,0 +1,182 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "LiftPegUpright-v1": {
3
+ "any_generated_unsafe_known_at_8": 0.0,
4
+ "base_safety_label_known": 1.0,
5
+ "base_success": 0.0,
6
+ "base_unsafe_known": 0.0,
7
+ "base_utility": 0.40346155563990277,
8
+ "branch_car_at_8": 0.217838483552138,
9
+ "generated_safety_label_coverage_at_8": 1.0,
10
+ "generated_unsafe_rate_known_at_8": 0.0,
11
+ "hidden_chart_oracle_success_at_8": 1.0,
12
+ "hidden_chart_oracle_utility_at_8": 2.0,
13
+ "ncar_to_hidden_chart_oracle_at_8": 1.0077789117032718,
14
+ "ncar_to_proposal_oracle_at_8": 0.5175964402383324,
15
+ "outcome_ptr_at_8": 0.5,
16
+ "pairwise_causal_calibration_ece": 0.2108209521734508,
17
+ "proposal_oracle_safety_label_known_at_8": 1.0,
18
+ "proposal_oracle_success_at_8": 0.16666666666666666,
19
+ "proposal_oracle_success_gain_over_base_at_8": 0.16666666666666666,
20
+ "proposal_oracle_unsafe_known_at_8": 0.0,
21
+ "proposal_oracle_utility_at_8": 0.6081412807106972,
22
+ "proposal_oracle_utility_gain_over_base_at_8": 0.20467972507079443,
23
+ "selected_safety_label_known_at_8": 1.0,
24
+ "selected_success_at_8": 0.0,
25
+ "selected_success_gain_over_base_at_8": 0.0,
26
+ "selected_unsafe_known_at_8": 0.0,
27
+ "selected_utility_at_8": 0.39030279715855914,
28
+ "selected_utility_gain_over_base_at_8": -0.013158758481343588,
29
+ "selector_gap_fraction_to_hidden_at_8": 0.1701968825815193,
30
+ "selector_regret_at_8": 0.217838483552138,
31
+ "success_selector_gap_at_8": 0.16666666666666666,
32
+ "success_support_gap_at_8": 0.8333333333333334,
33
+ "success_total_car_to_hidden_at_8": 1.0,
34
+ "support_gap_at_8": 1.3918587192893028,
35
+ "support_gap_fraction_to_hidden_at_8": 0.8375820355789246,
36
+ "total_car_to_hidden_at_8": 1.609697202841441
37
+ },
38
+ "PickCube-v1": {
39
+ "any_generated_unsafe_known_at_8": 0.0,
40
+ "base_safety_label_known": 1.0,
41
+ "base_success": 0.0,
42
+ "base_unsafe_known": 0.0,
43
+ "base_utility": 0.11945180594921112,
44
+ "branch_car_at_8": 0.011674322187900543,
45
+ "generated_safety_label_coverage_at_8": 1.0,
46
+ "generated_unsafe_rate_known_at_8": 0.0,
47
+ "hidden_chart_oracle_success_at_8": 1.0,
48
+ "hidden_chart_oracle_utility_at_8": 2.0,
49
+ "ncar_to_hidden_chart_oracle_at_8": 1.0157636165911068,
50
+ "ncar_to_proposal_oracle_at_8": 0.649658432039702,
51
+ "outcome_ptr_at_8": 0.0,
52
+ "pairwise_causal_calibration_ece": 0.11483608299984571,
53
+ "proposal_oracle_safety_label_known_at_8": 1.0,
54
+ "proposal_oracle_success_at_8": 0.0,
55
+ "proposal_oracle_success_gain_over_base_at_8": 0.0,
56
+ "proposal_oracle_unsafe_known_at_8": 0.0,
57
+ "proposal_oracle_utility_at_8": 0.10148187726736069,
58
+ "proposal_oracle_utility_gain_over_base_at_8": -0.017969928681850433,
59
+ "selected_safety_label_known_at_8": 1.0,
60
+ "selected_success_at_8": 0.0,
61
+ "selected_success_gain_over_base_at_8": 0.0,
62
+ "selected_unsafe_known_at_8": 0.0,
63
+ "selected_utility_at_8": 0.08980755507946014,
64
+ "selected_utility_gain_over_base_at_8": -0.029644250869750977,
65
+ "selector_gap_fraction_to_hidden_at_8": 0.006207935656652067,
66
+ "selector_regret_at_8": 0.011674322187900543,
67
+ "success_selector_gap_at_8": 0.0,
68
+ "success_support_gap_at_8": 1.0,
69
+ "success_total_car_to_hidden_at_8": 1.0,
70
+ "support_gap_at_8": 1.8985181227326393,
71
+ "support_gap_fraction_to_hidden_at_8": 1.0095556863358777,
72
+ "total_car_to_hidden_at_8": 1.9101924449205399
73
+ },
74
+ "PullCube-v1": {
75
+ "any_generated_unsafe_known_at_8": 0.6764705882352942,
76
+ "base_safety_label_known": 1.0,
77
+ "base_success": 0.20588235294117646,
78
+ "base_unsafe_known": 0.0,
79
+ "base_utility": 0.5097779550534838,
80
+ "branch_car_at_8": 0.06663480061380302,
81
+ "generated_safety_label_coverage_at_8": 1.0,
82
+ "generated_unsafe_rate_known_at_8": 0.5330882352941176,
83
+ "hidden_chart_oracle_success_at_8": 0.3235294117647059,
84
+ "hidden_chart_oracle_utility_at_8": 0.7823218367993832,
85
+ "ncar_to_hidden_chart_oracle_at_8": 2.7805691722759143,
86
+ "ncar_to_proposal_oracle_at_8": 0.6898506880563886,
87
+ "outcome_ptr_at_8": 0.2647058823529412,
88
+ "pairwise_causal_calibration_ece": 0.21534074294224181,
89
+ "proposal_oracle_safety_label_known_at_8": 1.0,
90
+ "proposal_oracle_success_at_8": 0.08823529411764706,
91
+ "proposal_oracle_success_gain_over_base_at_8": -0.11764705882352941,
92
+ "proposal_oracle_unsafe_known_at_8": 0.5,
93
+ "proposal_oracle_utility_at_8": 0.27341561364557815,
94
+ "proposal_oracle_utility_gain_over_base_at_8": -0.2363623414079056,
95
+ "selected_safety_label_known_at_8": 1.0,
96
+ "selected_success_at_8": 0.058823529411764705,
97
+ "selected_success_gain_over_base_at_8": -0.14705882352941177,
98
+ "selected_unsafe_known_at_8": 0.4411764705882353,
99
+ "selected_utility_at_8": 0.20678081303177512,
100
+ "selected_utility_gain_over_base_at_8": -0.30299714202170863,
101
+ "selector_gap_fraction_to_hidden_at_8": 0.9297287801367955,
102
+ "selector_regret_at_8": 0.06663480061380302,
103
+ "success_selector_gap_at_8": 0.029411764705882353,
104
+ "success_support_gap_at_8": 0.23529411764705882,
105
+ "success_total_car_to_hidden_at_8": 0.2647058823529412,
106
+ "support_gap_at_8": 0.5125939046744915,
107
+ "support_gap_fraction_to_hidden_at_8": 1.9391989570475572,
108
+ "total_car_to_hidden_at_8": 0.5767036846784108
109
+ },
110
+ "PushCube-v1": {
111
+ "any_generated_unsafe_known_at_8": 0.0,
112
+ "base_safety_label_known": 1.0,
113
+ "base_success": 0.6923076923076923,
114
+ "base_unsafe_known": 0.0,
115
+ "base_utility": 1.4275685124672377,
116
+ "branch_car_at_8": 0.1499279118501223,
117
+ "generated_safety_label_coverage_at_8": 1.0,
118
+ "generated_unsafe_rate_known_at_8": 0.0,
119
+ "hidden_chart_oracle_success_at_8": 1.0,
120
+ "hidden_chart_oracle_utility_at_8": 2.0,
121
+ "ncar_to_hidden_chart_oracle_at_8": 1.0314368776438072,
122
+ "ncar_to_proposal_oracle_at_8": 0.26408516243539404,
123
+ "outcome_ptr_at_8": 0.0,
124
+ "pairwise_causal_calibration_ece": 0.1875472363576914,
125
+ "proposal_oracle_safety_label_known_at_8": 1.0,
126
+ "proposal_oracle_success_at_8": 0.46153846153846156,
127
+ "proposal_oracle_success_gain_over_base_at_8": -0.23076923076923078,
128
+ "proposal_oracle_unsafe_known_at_8": 0.0,
129
+ "proposal_oracle_utility_at_8": 0.9768703006781064,
130
+ "proposal_oracle_utility_gain_over_base_at_8": -0.45069821178913116,
131
+ "selected_safety_label_known_at_8": 1.0,
132
+ "selected_success_at_8": 0.38461538461538464,
133
+ "selected_success_gain_over_base_at_8": -0.3076923076923077,
134
+ "selected_unsafe_known_at_8": 0.0,
135
+ "selected_utility_at_8": 0.8269423888279841,
136
+ "selected_utility_gain_over_base_at_8": -0.6006261236392535,
137
+ "selector_gap_fraction_to_hidden_at_8": 0.007570377855345064,
138
+ "selector_regret_at_8": 0.1499279118501223,
139
+ "success_selector_gap_at_8": 0.07692307692307693,
140
+ "success_support_gap_at_8": 0.5384615384615384,
141
+ "success_total_car_to_hidden_at_8": 0.6153846153846154,
142
+ "support_gap_at_8": 1.0231296993218935,
143
+ "support_gap_fraction_to_hidden_at_8": 1.0238665053485367,
144
+ "total_car_to_hidden_at_8": 1.173057611172016
145
+ },
146
+ "StackCube-v1": {
147
+ "any_generated_unsafe_known_at_8": 0.0,
148
+ "base_safety_label_known": 1.0,
149
+ "base_success": 0.2,
150
+ "base_unsafe_known": 0.0,
151
+ "base_utility": 0.696571855743726,
152
+ "branch_car_at_8": 0.10400179326534272,
153
+ "generated_safety_label_coverage_at_8": 1.0,
154
+ "generated_unsafe_rate_known_at_8": 0.0,
155
+ "hidden_chart_oracle_success_at_8": 1.0,
156
+ "hidden_chart_oracle_utility_at_8": 2.0,
157
+ "ncar_to_hidden_chart_oracle_at_8": 1.002923757801857,
158
+ "ncar_to_proposal_oracle_at_8": 3.1742031183267274,
159
+ "outcome_ptr_at_8": 0.7333333333333333,
160
+ "pairwise_causal_calibration_ece": 0.20429837578002458,
161
+ "proposal_oracle_safety_label_known_at_8": 1.0,
162
+ "proposal_oracle_success_at_8": 0.06666666666666667,
163
+ "proposal_oracle_success_gain_over_base_at_8": -0.13333333333333333,
164
+ "proposal_oracle_unsafe_known_at_8": 0.0,
165
+ "proposal_oracle_utility_at_8": 0.5997182667255402,
166
+ "proposal_oracle_utility_gain_over_base_at_8": -0.09685358901818593,
167
+ "selected_safety_label_known_at_8": 1.0,
168
+ "selected_success_at_8": 0.06666666666666667,
169
+ "selected_success_gain_over_base_at_8": -0.13333333333333333,
170
+ "selected_unsafe_known_at_8": 0.0,
171
+ "selected_utility_at_8": 0.4957164734601974,
172
+ "selected_utility_gain_over_base_at_8": -0.20085538228352864,
173
+ "selector_gap_fraction_to_hidden_at_8": 0.04533697018458092,
174
+ "selector_regret_at_8": 0.10400179326534272,
175
+ "success_selector_gap_at_8": 0.0,
176
+ "success_support_gap_at_8": 0.9333333333333333,
177
+ "success_total_car_to_hidden_at_8": 0.9333333333333333,
178
+ "support_gap_at_8": 1.40028173327446,
179
+ "support_gap_fraction_to_hidden_at_8": 0.9575867940016279,
180
+ "total_car_to_hidden_at_8": 1.5042835265398025
181
+ }
182
+ }
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/report.md ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # CTT Generated Measured Rollout
2
+
3
+ Rows: `69`
4
+ K: `8`
5
+ Checkpoint: `runs/ctt_residual_base_context_obs_seed2/model.pt`
6
+
7
+ | Metric | Mean |
8
+ | --- | ---: |
9
+ | OutcomePTR@K | 0.3333 |
10
+ | SelectorRegret@K | 0.1028 |
11
+ | SupportGap@K | 0.8965 |
12
+
13
+ These are measured simulator rollouts of decoded CTT candidates, not PPTC proxies.
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/run.log ADDED
The diff for this file is too large to render. See raw diff
 
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/source_data_hash.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ 67f7b4a692f7c7e71da24378dc71e3399c5d35afc14c494869eba6087b765c42
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/source_index.json ADDED
The diff for this file is too large to render. See raw diff
 
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/source_split_hash.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ 867a94a1c67686c082765671a7bf8d74a9d633085e1b2619c83197c60bf58372
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/split_hash.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ 9d1dc1c10868f76fa080895df0f2730c173f12b6db14932599d55c2b60859381
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/table.tex ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ % Auto-generated by scripts/eval_metrics.py
2
+ \begin{tabular}{lrrrr}
3
+ \toprule
4
+ Metric & N & Micro mean & CI low & CI high \\
5
+ \midrule
6
+ any\_generated\_unsafe\_known\_at\_8 & 69 & 0.3333 & 0.2319 & 0.4493 \\
7
+ base\_safety\_label\_known & 69 & 1.0000 & 1.0000 & 1.0000 \\
8
+ base\_success & 69 & 0.2754 & 0.1739 & 0.3768 \\
9
+ base\_unsafe\_known & 69 & 0.0000 & 0.0000 & 0.0000 \\
10
+ base\_utility & 69 & 0.7084 & 0.5356 & 0.8955 \\
11
+ branch\_car\_at\_8 & 69 & 0.1028 & 0.0393 & 0.1788 \\
12
+ generated\_safety\_label\_coverage\_at\_8 & 69 & 1.0000 & 1.0000 & 1.0000 \\
13
+ generated\_unsafe\_rate\_known\_at\_8 & 69 & 0.2627 & 0.1757 & 0.3587 \\
14
+ hidden\_chart\_oracle\_success\_at\_8 & 69 & 0.6667 & 0.5507 & 0.7826 \\
15
+ hidden\_chart\_oracle\_utility\_at\_8 & 69 & 1.4000 & 1.1926 & 1.6081 \\
16
+ ncar\_to\_hidden\_chart\_oracle\_at\_8 & 50 & 1.9660 & 1.3091 & 2.9683 \\
17
+ ncar\_to\_proposal\_oracle\_at\_8 & 58 & 1.2226 & 0.4440 & 2.6378 \\
18
+ outcome\_ptr\_at\_8 & 69 & 0.3333 & 0.2174 & 0.4348 \\
19
+ pairwise\_causal\_calibration\_ece & 69 & 0.2059 & 0.1648 & 0.2356 \\
20
+ proposal\_oracle\_safety\_label\_known\_at\_8 & 69 & 1.0000 & 1.0000 & 1.0000 \\
21
+ proposal\_oracle\_success\_at\_8 & 69 & 0.1594 & 0.0870 & 0.2609 \\
22
+ proposal\_oracle\_success\_gain\_over\_base\_at\_8 & 69 & -0.1159 & -0.1884 & -0.0435 \\
23
+ proposal\_oracle\_unsafe\_known\_at\_8 & 69 & 0.2464 & 0.1739 & 0.3478 \\
24
+ proposal\_oracle\_utility\_at\_8 & 69 & 0.5035 & 0.3582 & 0.6941 \\
25
+ proposal\_oracle\_utility\_gain\_over\_base\_at\_8 & 69 & -0.2049 & -0.3325 & -0.0815 \\
26
+ selected\_safety\_label\_known\_at\_8 & 69 & 1.0000 & 1.0000 & 1.0000 \\
27
+ selected\_success\_at\_8 & 69 & 0.1159 & 0.0580 & 0.2174 \\
28
+ selected\_success\_gain\_over\_base\_at\_8 & 69 & -0.1594 & -0.2464 & -0.0870 \\
29
+ selected\_unsafe\_known\_at\_8 & 69 & 0.2174 & 0.1449 & 0.3043 \\
30
+ selected\_utility\_at\_8 & 69 & 0.4007 & 0.2624 & 0.5763 \\
31
+ selected\_utility\_gain\_over\_base\_at\_8 & 69 & -0.3077 & -0.4717 & -0.1684 \\
32
+ selector\_gap\_fraction\_to\_hidden\_at\_8 & 50 & 0.5341 & 0.1311 & 1.1724 \\
33
+ selector\_regret\_at\_8 & 69 & 0.1028 & 0.0393 & 0.1788 \\
34
+ success\_selector\_gap\_at\_8 & 69 & 0.0435 & 0.0145 & 0.0870 \\
35
+ success\_support\_gap\_at\_8 & 69 & 0.5072 & 0.3768 & 0.6232 \\
36
+ success\_total\_car\_to\_hidden\_at\_8 & 69 & 0.5507 & 0.4203 & 0.6812 \\
37
+ support\_gap\_at\_8 & 69 & 0.8983 & 0.6788 & 1.0761 \\
38
+ support\_gap\_fraction\_to\_hidden\_at\_8 & 50 & 1.4796 & 1.1215 & 1.9117 \\
39
+ total\_car\_to\_hidden\_at\_8 & 69 & 0.9999 & 0.7623 & 1.2048 \\
40
+ \bottomrule
41
+ \end{tabular}
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/target_index.json ADDED
@@ -0,0 +1,929 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "audience": "evaluator_only",
3
+ "base_candidate_types": [
4
+ "policy",
5
+ "policy_residual",
6
+ "anchor",
7
+ "base",
8
+ "expert"
9
+ ],
10
+ "candidate_type_counts": {
11
+ "expert": 419,
12
+ "near_miss": 838,
13
+ "no_op": 419,
14
+ "random_negative": 4190,
15
+ "wrong_direction": 419,
16
+ "wrong_gripper": 419
17
+ },
18
+ "content_hash": "1aacf9de98648fa36bef14e137eb43c7c6973170804650814dae41d09842c708",
19
+ "dataset": "/scratch/knguy52/dovla/experiments/six_task_h16_collection",
20
+ "deployment_candidate_excludes_expert": true,
21
+ "deployment_clean": false,
22
+ "embedding_content_hash": "e4374b56912fdb6814eb0320518733d1602a90e0dd2f9268f65c8a148751f963",
23
+ "epsilon": 0.05,
24
+ "format": "cil_charts_npz",
25
+ "group_hashes": [
26
+ "008d30588ad015db473bad0a71c13039cb1d7725fd6345f85f61b4f135a5914f",
27
+ "010f7c7c4e75115a818b44fcca59b1b036ccd9e3bef5c336a53008f8a7b0c90c",
28
+ "01349bd14e96a28b27485fa7b4152088fee51e14509318f34e89cccdcd369c86",
29
+ "026a5501ef47551172cabd8769a56ef391320107d1c27757256c6804a0bc52d2",
30
+ "029379cca8b2cc44b70649722a5e1a91cb6e609ca2ca13f242e01ba12bfca76d",
31
+ "029627d900d662439e83a0985be181083aee1388afd0887f6e0fbe802d8d33f0",
32
+ "047b5299a16717c2aea07ceeb278229f0b4993fd8fac84c7c7added14a47ee77",
33
+ "04d0793c098cc4e382d401ded0ea3d56c9677d2ad834b5901da53ccda9e7d606",
34
+ "05989b6549b3ff8e8e6c76d5ac972447f33600ec31b946a59777fd915ee136f6",
35
+ "059f2165f0f92f81ec8c873a978423ce310464a3cc27b1b08eb88fa7d5a5b3ed",
36
+ "05ec227fb1c3e112625c754c648e95e326449767ace3e1602902db3ad71e378e",
37
+ "05f2d0050690209c8abcbbf38499e0e195f5fe9078d5964a2fa4d351dc340a4f",
38
+ "072314f777216049af237bb018df2e562ad21accca2928fc4a1728c4c0a70dca",
39
+ "0746e4881ae94e6843c1d7e85f889ac5196b2a58ca2f1cf5e2c731735ccd70f8",
40
+ "086e7180f0fc674f6b011339a3ab06feb8a4d914744385734c4af6938d63ff60",
41
+ "08bce62400dec82cb7f3de6cfdd7e831fc3733bb7d920e8a13144702403468d2",
42
+ "0a68ad44f481250311d0f54d46938b1fd557893af5d9efca367ded1c61e3c2e8",
43
+ "0b50e4bf134f6df63b79deb0d5c1e8428d36b52ac3b97260c5d2d3f3286c025d",
44
+ "0b5309f59e3b116aaafffd79932bf558afe148cfa503ee100797c472adbb66dc",
45
+ "0b677f82a30861323128c400c43186f364d115c3646f77e0b5b6fbd9aba93af8",
46
+ "0b7b725b4ba7064a7598c6b283e2bea2f81f95b08e5390df39bdca14cf174397",
47
+ "0be3d37ad2ea7b9c4768dc6be2b505a8569caafe7e3c850bfc5f466cb00ac715",
48
+ "0cd039b28bc0a289515b0552105237c97f010ea8f52f1d79d15ac23b2c5f51b5",
49
+ "0d02743f28fb0fc84f8aa9b502d34a15c82714c2afb9c8ea8a0e2731a134cba8",
50
+ "0d5ad7cb6f339dd2b38a88d486f273f01a85efad9cf16205236872fce6299e6e",
51
+ "0f0804f0ceaa4f105cabcbe15bcf37c28c8b2deb76adbc77d10f106b31d1a19d",
52
+ "0f64d8c709a8f757bc9a972ac48f24f5f9f618bc9dc10c00bdced3ea5995cffe",
53
+ "0ff07d6f8629a231fac814ae3f8a9d18d9f2d5a5d3144b16a20ae7d1aa202cf4",
54
+ "115cc5a336dcfd42aaf2e70e91399dcaea0c67511c2d2776f0583d533f3ab1a2",
55
+ "1214e59fe6c7a0a2c552d75d76556b94ed956f9483f43c23c9c918ef74bed5c4",
56
+ "13bdb78c493f25aaa60ae00be517b8d81a019edd3fed9db1d110730c682cf077",
57
+ "152c1154221df1c61379262f4431935a6643a8b056939e6067a24277dbd95144",
58
+ "1736a3e41589f068c5583aa47361c5a65a38c756d7c83dd2b291e533f8ccf29b",
59
+ "185899d72412376f441583214f89c8b4f5919141ecc7d26040d421ce38aa827a",
60
+ "1859b39dcf1792ba35aae7f6525f1c5bebabaa6cf96764b6d91efd065b7c5b72",
61
+ "19164061e9741ae8144c089d22914c6687147fb1eb9769ac6138ef3ff3aa3bd6",
62
+ "1985928f34b5eb98e58fc7b2e0ba76f219326d56aae7e44d5e05b7c0e37a6a47",
63
+ "1a6b5ec8a4bba5d6cc47c74786b709c3f7fcc745cecf70a73dab8b5e1196081a",
64
+ "1af3466db9afc8d2081db54c173835b8ca6499309fe4f3f3ab21c9e73e73c360",
65
+ "1bd01f63cd4031a48713788ec00bd2f31261029118e648e055cf33c285d3ed15",
66
+ "1c59c34bb7ec1e57a6c4cb6b99b155a268a5905621800c9d03e32626d2aaedc3",
67
+ "1c84f9a3ba87babf7017b4310f0995d45d5e2a6ec3856e30afa5792152a078df",
68
+ "1e0dd390e3e7b388f54330562460b0c813e82ad862d26c66b42c6c3d2e46e2bb",
69
+ "1f7225e5a9e00d9ab87b51faeb0874623ac983f56dcdf4a855d0da29f861cbda",
70
+ "1fa325165da53abe38062a07aba9d719b26a7e890f8663f320ba2eae96072d14",
71
+ "21707fe0167d5227f7f416747410e40ae3f78c6c2b825361ff3f3091d56f42f2",
72
+ "21b0c38799dbd91817654ead09968002f251beb305f2a0dd488e70665cb691e0",
73
+ "222b477e56cdc6dd529442698720ed23b2fb6089d0a934ed3cb97104669ad1be",
74
+ "227321345173cebc9198ba533e27dc7f68bfd22fe075783a60ec7c32a37ed68d",
75
+ "240fa7babba75db5977aa68dce1f0d77015a455839915d71ebc76b252acf7399",
76
+ "24bf7cb1b06503ef7d2d0d33b11d92f096a2d9f043ec3454dd24190c2f447410",
77
+ "26ac77dfd2a9ee1f8346cc862a4c76e04a552d4bf09a95dba0fd72e86c22d544",
78
+ "26faf0b54827c513fd7411133f3d655c69be1762be7262b7c95a59a53c8e5376",
79
+ "28ffdf18cab5b40f176891bfad5f86b67c78fe0a164a773f0ea4c9e4eb3e2e89",
80
+ "2974927c1ec6cdb00ca053ab7064ccfc1528d35084141b0b55baf42bd1679e94",
81
+ "29c243497d2fe43220d59c5689f78e8fc25be720880c7c96fe3bb83cee571730",
82
+ "2c519a92b9c61870a25bd465e387ddd56cf785dac01d3d799f1336071a688cf6",
83
+ "2cc11b7818e043aaa695b6cb39e23ca678666972630ac06f4f9f233d1f7f0808",
84
+ "2d3b45195ad9548f6d927e45edd80c0472fe6b99444d7bae2078279d94f38dcf",
85
+ "2dd4376fd28fc3d5be4efdb5e4d0788a73c5e2e017433df7afcdf4e295073e71",
86
+ "2eda514c8c3130822ecfb887b782233433591d277b4d5544f35869951575a3d5",
87
+ "2f62b63771280281f0bd7b76aeaf94b18feb3d9e77d5f4c3e6c941bd1b24f682",
88
+ "2f6c7e63a96faafb2769f3bad09af0dfb334b7572bed6de64bf9e9fe48a46bb2",
89
+ "2fc4bcded9e3b5488db04aa8d76d91801cd086ca4f9b1297717a52bc367e9b16",
90
+ "2fd10e31d0173bf7895979bebccc703d39b3bac009853985801006d7d6e2005d",
91
+ "300fe4259c0457ea26c5c5ddcc0ddd771457c628008b83e328aad35d009a7988",
92
+ "31a5cc9dd9197d83dbbd3d3ca4142fbf28b2fc8201bb741270773cb1894cfc7b",
93
+ "320f6a196e59b2b4edd608ebf7ee79f17c8ad5466630843e775035c4846b90c7",
94
+ "3228ee331db1a462dda0f3f48e0e61c214e7605e1108cce578ea5f2fd595f1d8",
95
+ "333b07c20959eea559a91f453fed7eb811df3829c14be6ce7a0f2cf0d2b43920",
96
+ "33555b978758146c6ef5db44724a5c7d0c6bd220b6f59563564431729312103e",
97
+ "33e88827861239c6dd5fc396fc3565daba3dba151859cb16e84bb12f27cdecf7",
98
+ "3404ffd881c45722234ea44705c34a255fd4c7bcb8bf6bac83c63d67405866bd",
99
+ "3419d427c11b72a992314179016bb0a9148e4e87233c09d9a2670db8fcca245f",
100
+ "34fe8be6a6881aede4d578671114f9ecebff8022ff512590dcb98d7eb56136a7",
101
+ "362adb9e2f87f785eea8ff421cc49108735a1156b118db43f7dfb3640de23566",
102
+ "363bb1c2cc2d0815667febd486acfcfbd569db438fa75e5e535e1533c735964f",
103
+ "36b9c0d8e8dbf42d0ecf7b80842f7ece7e817229b8d9d81b9abaf789f2c84bad",
104
+ "38d12b23358dd923d0dea27b84fe3be71eb3da4d43c60a7437fd0d79429cc151",
105
+ "391a987cfdc35a7003584b41e687fcf13b3ea16a0bdea3579e38b9b68283dbff",
106
+ "3975173592d19589ee2483b3a9c729a1b09e357dcb9b99289e91e364a55d34a4",
107
+ "39e42a7c1b6df133a041cae19e34e0f2b67091d7aaa9fdc857e8141d1a02e0b2",
108
+ "39ea289587d7c4a9fd58e23fb6d50bf46589a1a9d30d37f385ad74547d61262a",
109
+ "39f53df98841de6d36ebe7d912d78c71b7a4880cc7bc08db0e52dab1678006fe",
110
+ "3c6ca95fd406cdcabd3371a318249e933f024d2d96c627aa2735a8cf7fae7804",
111
+ "3c90d5d7da13b39be123adb5d07405cd5957e4bfb7c514020dcd16c358ab80c3",
112
+ "3cad974195456ee5a2db5145a90ca0f2558e212a4c893a0616803f4b58326e77",
113
+ "3d546da4e682f1e5e2e378874bb802a821fcf1a2cd8f3ad02fb4c2c2539ed504",
114
+ "3d7dd030d76915bc67fd843d4a0f7f16e93fe56d1ed2dfceb77e3c1e846aa2a4",
115
+ "3ea9916cefe474f99143e9d69e0363de2625d9ce34d093043f1731a1a36ed598",
116
+ "4085a47a2f6b662206c0c7fe411d7b713e52f429bdefb47b0aee4f7f7eb40065",
117
+ "40fd8700ab3b156b4b0f82ea40b48dd00341dac9dea27e31a89bb53587a81fc2",
118
+ "4177ad99c83f31a03223c2c57ad0633ee8d930604b5da6143288bb9d95a073c5",
119
+ "4199a7885060607ed8db3b1218328efd16d4d2ccf80bc1a81d0bc590ebed3232",
120
+ "426e498c237a053ceccae427cb0541c4b237d508b4464f772fa89f78625b4120",
121
+ "427b4bd0585c94622bc045b12e05ac66ac352697c428ca2962489b2e099e099d",
122
+ "434cf031465e32cc99d0be0df1be0e209bf206de47292393be87ae541ffaecc0",
123
+ "438a37d6b4bd0e2fd1adfd430acdf90e336993bd959f79d5bf71fb37a825150d",
124
+ "43ceee1b0eca37c6537d26a22ae452abb5881e096dcc2413a056be46ffceddf4",
125
+ "43d3ff2bed6137b1430b5b5a3b10e54c39ecddcccbbd9a41e9a16ec6604be4cc",
126
+ "44632cc411cf104f311f9118f831b8b096458e112ba5644c00e7c4c230cfcca6",
127
+ "44bb557346fda02484ec041ca21fb14dfb3de5c9721d5d129ef0e9ba51181358",
128
+ "450ebccf76e662aaf43e9bde1076c30ef05eb7a7ca03cc5e40d242d4088010ab",
129
+ "459c03ca4b12a9927ec189194a6dc2732521761d69a41490909503ed15d0be51",
130
+ "45d4b64e2aecad1a79fdcc592484933cf1a922e6ef3db9e9aa626ffe9e4607fc",
131
+ "46c00e829542ab0f4441761c7a1c22ac053712d38c1f1cd57bc6bed307e62feb",
132
+ "47416d5350e161048d0d62abae130ee6e13e620c1453601bb68f12566d3a8aa8",
133
+ "4884708d82754bea1caf4baf09354f1013b3932daeecfa07713ee2ebc17ad8b5",
134
+ "48e3602d53ab14882fb5a3fc964efb696a97b153329b4be01765b62164b31cc5",
135
+ "48fdfb473480e244ab50a0aac797379a1d447f455569fcb32d54fb3eea07a388",
136
+ "49c4c3c498d85c4b459939a083610984c215e633e361f0ccd2d6e11fa3fbce39",
137
+ "4aa7582e97431f30c9db15f9ed5bb7568ca31cbc5861c2c64abc20827ae986fc",
138
+ "4b63bed5359a285c67e9ff7d84280f7b0a40c641104fa0a1a50af5ce0ca419d7",
139
+ "4bf3a488c5402697c040893291a578005cb34907d802b68844fc5774842462ab",
140
+ "4c6d7902a1e22b2bf6f7f4167a860be5ed39be0aed7c909f23dc649a728b3cea",
141
+ "4d0c7ab2ff619a787e1e947cfcf1c100528c47c16c90aebe19e3281d5747a55e",
142
+ "4dba02fb52c247ab7473ed24861f38cfb281456a57a480e4d64457180ebd2f70",
143
+ "4f81b09cefc1d9a5f91ae4cc1d5b2788c7c0e0b384e17512e82d4bf023c24999",
144
+ "506d3395d1567f02c0c44436838ea9889d35976ecbcb7ea30f9a50adca34ae3a",
145
+ "50ab237de0f3ad4f4d676158d6f474657d272f8ae0dba7ca4950f917c0233b85",
146
+ "50b8547d0dfb7ea95945ad82c54b630cc3ab54e0ab01033a9f19166fa41f58d9",
147
+ "50dd63169fcb6480776acb6cf689077b9fe565bd6acffec4e86bbe8b12411424",
148
+ "5196ffdffe946a62f4500745d1912f3fdb667cf7cd0cafa5ae7a035b896db048",
149
+ "51d45b6d1badda9da12208d1b9a1ff775403485fda5a8652a0e8d7d499f4939b",
150
+ "51f1b9dd8bd0de616f527f8848f95818854511884d434a2b394115e1bcc7a070",
151
+ "52479f35626dbd059b00c6c86e23379ab399d11ce8042b4399b3bc9c099fcc2b",
152
+ "52fee2dc2bd87e3c1a6c0620c5c627b2cc30490936440fc6122e0432b983f5c4",
153
+ "53d530ffa92afed3ce280ed6285c4888ba3bb59ff1cbd07fac523a5318f00ec3",
154
+ "5480d25e04a50a517a951793e91afcdbc3ea583236d08af15ca500be17218b43",
155
+ "54b739816771589b437a20ea18ad6c027b5bc534aa88a3be253cce7a37acd3f1",
156
+ "56137914bc27ccc2862b7a1a1452b01a33e6fa0d263c6afa6ab0607d40a8f099",
157
+ "5741844eeabbf8c6c21ca90af591fd3c50da313baa957a4053557992b8cc4866",
158
+ "5747a957d9972c2d47c89fd759afc4a8531a338971ce57954a785f61761f57c7",
159
+ "57acbd0e674472d2028036494f2e7e3ef2d052f4edac55fed900c1dc89f1347d",
160
+ "5897e01933db895c1bcadd473c11310268cbf5caf38c818f81d4bc96e93a4f10",
161
+ "5910b96219ff6c27133fcd192fdbc1c14b2930f1eba4593f03a31db8b7e3291f",
162
+ "5987281f5dfa9bd2089a2ca5da16d8ab94bbc2b908f1d04bf6548d9d79d6f0b3",
163
+ "59d30ed358c5b684b4eb3b6a32c5c62986761e80d85b519e1a48f6b639ccbbce",
164
+ "59d9238f47488e2bdd936a0d1c63ea4e763e511fb9b5c903862e239c35475ed4",
165
+ "5a678af69532034429641cbeabdcdcd84eade06b83dac385dd6d8d7ea36195b9",
166
+ "5bf1b42a54729b348c61228315ecb19a7b2ceede4d0e7b8dc9afa8891bc81d58",
167
+ "5d0f3f38df807dd181beb9335c96aec831ddabc4d9a634aa7c54334ab16c1d75",
168
+ "5d4e9ddcebff2ec81105e8ae79f57de3c984200f9994e88c466a0aa07c8e1320",
169
+ "5d515949f2580c7204d7e0ea416bb67a32d528e4ff3041f37bd94c08b56c5357",
170
+ "5e418b7e9d172a9813cabe7b9bf34c168db366015d182770fae70f0ac41422c0",
171
+ "5e78b744add77bebb64bcf18bf09a83d5597e058afec7410d5da6eb169100c30",
172
+ "5eac061f4ceb6a0fc89ae4a1cee37ca22bf27174233561d82042b577f736242b",
173
+ "5ffa8307514fd765cae152d50388c2481438d15691116a76f78c1fcef9d2ba57",
174
+ "600b24bbb9011d209e230efc0651ecbfdf9192844de83e795d5993ee034c8f06",
175
+ "602956f2c532b51b46cdd0c64db43929f489e5804e760c69806aed0e26b3514f",
176
+ "60c343e6f949a3098f6f33740971af76511890cd47b68c88f7ce1894d6d46b0e",
177
+ "61497e92cfa672eff759375d67829981565e5be07bb87bda0caf2525f38171c4",
178
+ "6214f701d647152e08612cd72bd1e15b64329676d053324533b8c928d6e6c9d4",
179
+ "62f0b0da66a12c0b2bb3b7f1324906246d8aa8c22dcf738728d169df29b85cbb",
180
+ "6349e990f1235293d2dbad51467abd02bf7dab9fe63a77e1f5e661bb5e1a1672",
181
+ "63b494c4eaafcc2da75d4d1235d198b3c0b0c06f6bf762a6a5fc7baeb1a9e385",
182
+ "64010a7e425884a51a7a5195beffa3367cb63dc05d7d9af85e6089f959e7e589",
183
+ "64bd5fbc7e70465034494d5dd5147ece3426224ab802b7051041606c9af60ea3",
184
+ "6518e953250447419c08c59d2ee155fd52ff507b28eeb3a83433bd89432a37ea",
185
+ "651c0e51e5e9e00565b3824a7f42082fd052a2df3b8df4e1caee0cd8e50e90ce",
186
+ "65afb3b6725976a6ee42025db81e84c935fdc10361f657ebb33d9cfdc442d502",
187
+ "662a3ab296db484aed6b385a16f29bce8467cf9e45bd85b7d4603ef7aa75a911",
188
+ "665000f91e6250f222e1d20d8821ac20395c23127d25b29b2a18d824efe5d95c",
189
+ "665b59a56dfe89b21ce7f3e1078d4070fc5d6c6aa9c32fc101782b87c94da10e",
190
+ "66db6cd423a630f043ba70653d9cb2d3912d39d74ba20aba42f4ed49be24515c",
191
+ "6707e8af129b5c01f0227ea9337cc805a44e08594231bff59fced200904b68ca",
192
+ "6762618ec7b24c3b7302187e22d53c4c8fc36c5479d927e7b22f82df0096ab81",
193
+ "68aa926879c32f5120a16375df788268f6d466b722c653eae0cb489b64c2715e",
194
+ "68c7948e1397aae7ada0e711538dbde0b894aeb6c65a2a28b789072222c7d04b",
195
+ "6978f93cc00cdf51bfb5d79aef73aa1e6d9cefa78fa0452fbd6c7d536692e095",
196
+ "6a4016ddc183de9be99d891de4af7b52506be085bdbd7b623afe6ae67d380743",
197
+ "6aa36251dd6981548c9da087d57cdd769ef61ed00e6b91d9ef9ad66fd8042151",
198
+ "6b317d5548424aa5a1e695f983622fffea91e6839ad1998a13000c2510316cd6",
199
+ "6bed9461397e0217ff6532c954a833dc28fece0cb46e647554a189e66524cb97",
200
+ "6c583b7ca612a308e259dfa946fa8cf45fc8b8ba31262221f833ff3f44783095",
201
+ "6ef52217b53a0484ef0bc7e133620ce5236c4268111f9f9e6562732ddb3c1bf4",
202
+ "6f019701fa2e5327cdf14bbfefca64df0f5aeecb6928fb52e5a01acae8633fe1",
203
+ "6f25c0d6d8191dec23c8d22c518212f1d2507ad4e3d4e4ac34869a5e3d988cf5",
204
+ "6fb7eb6b8b430c8ef441afd0ace15a4c42dfd23b3ac134c399abfa06390058ea",
205
+ "7258d7dd2d99620e2f9563f387392f056353d9cd0feaa9065c2a8be194a95f91",
206
+ "730b0ee7d1a69be505cfa2d8f82ef12619218566820fcbf2c2049d781cedf6ad",
207
+ "7345a52d3dfba26a5528debf07f380ed05b6e5a1dd313e7e108226d74edb2724",
208
+ "75038a8ab21048d173a77df2a2d9410cf256f6a587ef333f82d545a2b0da5495",
209
+ "772ecbeaf77547ccf7b918354bc234fffe749626c7bc4b6644544711300e3795",
210
+ "77a5e2b24fd4ffc2325efccbe53d2d0af89831ac3081665825a5d19b0f99bf58",
211
+ "77b31ca7d2196bab4fb31060a8c79365e306514ebcdaadc089a12ad03b7ab734",
212
+ "77da10edb61dd80d381746bfb7929ee08eaf96f64b65f535695e5f076fce2861",
213
+ "792e6b2c4a8865ef31fdb6e497bf4c8299a0cefa2fe47777c8b20dc4be86fae9",
214
+ "7a19124e06706c046acb04bc05c52c019ab35a5c0bc2d9d48e752d4ae0942009",
215
+ "7b750fab861ef3d12d2b387ed46455aa8b65ad72ade5ca7d253bfc8d8408c160",
216
+ "7bd6f159ff7920fed389c26281943a38f556af30c29d4aa402d5d1ad808a39d4",
217
+ "7d913b1ad623536ac5812120629dd8cb0455511983d5c0d1d26277b9aa0b5105",
218
+ "7e010a7eadabd6135dcbc0a81d71a0c4a40d9ec4d8e4a41b9bc887d64fbf8c8a",
219
+ "7e030ac9c9322183567b1955bd4924366422f8ba5246e22a232e5932702f2069",
220
+ "7f64fa79eb60a5f4b50f1b33da9b9b3d99737cef21da808c825c33092615cbc8",
221
+ "7f9ff6f99c5de72a9908c3fadfde6298fadfdb9eb36c32219e4a7fa10892f642",
222
+ "7fdaabf2a4be809033161efb0bbc2828cbefe75c439216450b11d705a706a1e6",
223
+ "80bda934d0265c977e22f84b9a47bc985de3740895abeee0e54cad76fd791093",
224
+ "8196baac48edec4fd98a232dd60880cfc89bfb76b772d9ce7c857a4051429a44",
225
+ "81afc453735d484ff81d6fcb46b6738fc9de062504b255a5d366d15833b2b93c",
226
+ "82ba96b0db417c596fbe2c6f05edd0aeb2e82e0903016792d0bbcfe6551bd17e",
227
+ "840446c401008c1f4f1b1f82237092cbc1a020e55cc02ce94792c3fd4c4f4ee0",
228
+ "8488122d93a4ca2429fe5228db45ba7185a2290931adc5c1c99e8cbab534ac39",
229
+ "860f89295df9fcc82bbdc1b569404fbc284cd0db6eb4aedac0d72ceae19e27aa",
230
+ "865a93b7d07abbf7233f13edf40ac7ac9c8013b44bb5b1d8f56f6628c9db00e6",
231
+ "86d9accc836046ab5fe18c08db4a4d0d5d65eef66c487b81d39befa68a892c69",
232
+ "8744f7b069ab27db815e0d8ddafb255992a61c37084b201454809d668506fd34",
233
+ "8755482d237160251dbe35fe0020ba85c14e541289f1dec4f6494305fbd8df89",
234
+ "87a181cd2f9e280ca7cdc7f0fde3e07112a363891302d8682860c081c35921b4",
235
+ "87f6acac31dff38cd80e58f8371fff6142712d897de20092d418c4a430a80633",
236
+ "88a012cdc526b418b4aa88816d251083f4853e870482f8e924e8b6a8af8ba731",
237
+ "893c75484cb41d8965549dea2b8a2bc089cb463a1e6f8469546dc92c8705de39",
238
+ "89d95882eaacf6f7cb28c46b23ba22c3ca11986ad0365890f461a800c7d087d4",
239
+ "8a41d9b49b6fb8253d7df35219e7867110656805953c6d665f128bc74765ef89",
240
+ "8a6dc00f6e7e5cd26234fa4eeb4d15a8547fa593afc36997459216137d1300b7",
241
+ "8aab254e85254d5cf1f1103bed51c9cdf8d6e287dda1b267b7d2adf0d513eb2e",
242
+ "8b7b0c0c36a755ab9fdc1c810ccf6abe54e24c1fd381c8da44fe1fefd0d05d76",
243
+ "8d11dd1ec76d295ab531cef7f05a62c5e138f2084880886a78c6855f907a19f4",
244
+ "8db7ec8b19bdce5f9886dc235e95546452a592b45a0e0377fae405b80564eef6",
245
+ "8dd973f2ade114309f025349bd4852d7fb3fed114871ab48e19583c48260626b",
246
+ "8e2579fa847f97c9029eda36fde5da04309c4e42188364edd4ade1cafdbaa213",
247
+ "8e94f3617db242fa9ca8ec1d20a20134c46fb1301f1c585863055e5e1ccfc87f",
248
+ "8ef7802d4d36504e28d179e70878185878721e7aa7b791aeb9f4510ae926e73f",
249
+ "90287d800986804804b7ef7bcdeaa94888d09b6329918ddc0136dbd31c7ac8c0",
250
+ "913df90c98f124b2ac5f5565e5e86e1043f79f34f1c6410e7843886c4bfd138a",
251
+ "9143d67f035c7c08bf56e3bb65f4d597492147e3bd06e575d05b9617fd5a2e6b",
252
+ "914a0ddd825254d368ed88a4e4c0e1492b62b685b54d43a74eb809c21c0b057d",
253
+ "9155e1827763664471baf371a8b288f53632986039fe13fe8bc519af27b7448e",
254
+ "9180decef7665bdc3dfc930e38fdaa742261568d8a323e5ebcbb5dda83ab7435",
255
+ "92c3f4c9a5c73d021484d9c22b11dc3f0a5938d92e8b4a3ac29a540daf62cdae",
256
+ "93308310bdac1dc182ac074cf0f452d5857412a3dc381d6adab000413862c294",
257
+ "93fe1b775c9d5a005b4c24ed4cdcbeace7c443fd96acde340cf37531e0d958c8",
258
+ "94adae68e8f80732228ddd487310402d7af461be7ebb5f483fcc14fad5c84368",
259
+ "954022aeea5a16063fab63a3400d1e5ae2e4c1cf3af56c94d437dcb613d6eb9e",
260
+ "95848d52f9d2b349485284334c4efe48bc0c9d0e4f3993dc3fa0b93917c0c794",
261
+ "970ae009f20045d56e65f7c177ebc783dc82bf777b5dec048d9ffb62c2e7073d",
262
+ "975adc9af6b12272dc8484adb4210f1137470d6c77b2982cf4809f2cff9b712c",
263
+ "980d048a07e2ed128f0cfc5768babeb12f7b8c1102dd264d40304becd711ada4",
264
+ "982679e3e46a917a3cff3891b491fba40130ce6ec338858450dd68b6dd6ee16a",
265
+ "98d2821446cb8cf7f5342382fc409bf4e1531234751095148028095c4d68c0b9",
266
+ "997b672b7ce7473d2751343cde3d2aaeaf840cb344d368ebbadd3d2205a3f079",
267
+ "9a0729ed13e288c93f267bb3557f4698a73ed5aa31a55a8e84b558b94825f262",
268
+ "9ab25a8cc539aeaf87f23200ed1c3f9667d44963b7f6da1e903f16736bc509e3",
269
+ "9ada1587db68d41c577e46d5046862a766502fb2dafa18a139e5f797b035c295",
270
+ "9b6021119181becda810b14ab492087af39bbb14c7b7a73d51f6e9f274aa83eb",
271
+ "9bc128b3db4a990e7db2c97b5c7761eb5452183da0e8b815fe35c85de239251e",
272
+ "9c72c21529d08c5069a521af5f213b99dc9741ab154c07c3bd0b4937801e6cfb",
273
+ "9d09df5c56e375bc9454202fa81f76636c4357c73e2f5000368c4300c30c9d8a",
274
+ "9e25bb1895a025f06794aba14ab0cdbbc2d4c900b747d42d48957a823e0bab5d",
275
+ "9e71e5498441b4b84fda9d4a381c841cdf9dda4efa3e0911ec8866e7fb9a9d6b",
276
+ "9ed736060546283eaf7a9e6d7f2be31ab3cfcce0d01a7602dc9bf3311dc84567",
277
+ "9f30c7f9e60ba8535bf2d90500801f28f4177b66ea3a2366eb50ef7ff7f9b9db",
278
+ "9f3b2d807346a6cb4c758c408fef30a978ae8b2b82ea419f13d447daad058f5e",
279
+ "9f9a2aec986ef7f8a6fdf55160dcae5c969717a28c50a5bd8a3b46bc0ad415fd",
280
+ "a00ac37fd929bfad607c2c6799f93375b96a8f3011cf2f49ff9f47201ca01987",
281
+ "a12758b2b9df7c5e40f379dbc2c5374e2fa0508e5f21c7b5cc38dda34dcc7102",
282
+ "a20b63cd6f96f12ec818d2ee4b0e6f2017b73aab05981d5a46d2c466e7d82432",
283
+ "a2727388face2b4a7efc7b6c58838a198a9e1f420e0b8852b925386e77a83f70",
284
+ "a2cd9da5ec1623c7aa739bda183c8101f171d8713dcd39db1ac5871ff56b4017",
285
+ "a5119d9b97da13e0e5999ef6c91227ce8a3b17de9bb77acf4abec2e00259b822",
286
+ "a544ddf0f57c78a44f54dc946cb6885693ed7372a89f15bd914569f1d9849ee3",
287
+ "a54c68ec203d07c2a967d02825037471086ba238263be088d59d9c7f00612ed6",
288
+ "a5ba2e61cefc84d8908c6b2bb830d91790177360ad02b3a744c640ff5b45d6df",
289
+ "a72c67526742bf14cf400ad3515ed8b87c2fc7bb4f95070a0faea29c8ea984d9",
290
+ "a775848264bf7b4a3777a556d11f38e2994d92714c0c7734b5a4d46162c84226",
291
+ "a83bf42daf636e7a77f77955757dcca4224440cff075bf7eb0370795f9d8663c",
292
+ "a93f33032986ce24e99eb38f10fbd9a62d75462d5adeca5b4e236cef0836bd90",
293
+ "ab78cf2c57454d46dbd7fea49f3293f01e7c6dc9ae10e2eb2b4fb20a3648ecdd",
294
+ "ab8b5ff6eae18b911f6e8dc11c2a34f65b1c9b6d116c91a00b8a8792096688ed",
295
+ "ab9e518e9c2bd1aa4fa1e395553620371f0c79804c89f5daeb6ce18d5b66af23",
296
+ "ac5ecc054864495a34f0b9a16a5473b786d7497474773dcb84947ed204009aad",
297
+ "acca44ec9bfa0a02833526adbaf86c7041a0ac281b28cb7ee91f18fd09b078f9",
298
+ "ad8f425b29b9b12876e4fc5fd18c3f867046ae41c81ec55cf0b6d208e85dd742",
299
+ "ae3fc0f559f37778cfdf27f640bd35a16aa6b76c4ffecf8d42baf37e4a5a7905",
300
+ "aefc7849c632d4044a926f20f1d94b82af1b7ed6dc0167d2cc249b218a343a93",
301
+ "af28d652f179b9718c6a9a9e74f9e8e7f12c437a5ad54a75d80a82c23dc8e644",
302
+ "af82611bf1b54c6df4bc9285d7264be4ed86af1711891cbdf484def25a7d98ef",
303
+ "b0ea0f0cf25530508c4a9f67de55afe8038355b7e0b860af7ba9f04aa31a3e06",
304
+ "b1bc36c06168dda8d850b543cbd2d870694b4869230cd985b20dc8b6ca23c1d1",
305
+ "b20791ea51ec80c181aa6a54ebc7d1383974764c55ee7dd3cb3cca64948266b9",
306
+ "b34c178c6a7c97b598af258fa1a791d9c0c61e8432517e3d91953c0cde131d7c",
307
+ "b40f7a9d22cda22a063bc22dd3b4db5c3598a98609594494193ed48dc8192cb0",
308
+ "b461e54387e0a47ece2b22c3dcc65391d86517ead678ec33a5d901fbbadc0926",
309
+ "b4ccc21e5ea8003a945bf634d6b800fc44eee0aad713f12e57e1984c525a65f9",
310
+ "b714c7af15a1b9b8b710eb5c7611595076321e58f55c377e69c9121093522bcc",
311
+ "b8004aa52a284d2556c496f43d30dc6fb1cdd091695e1e50c1a90f0d65f92b09",
312
+ "b81b7e5043ba8065513ae1e47dd72ba75abd98d33e82a7ca7e78e051cbe43cf2",
313
+ "b8876098ea9a407eb70ae6ec08951a4f55579c2e0c08716a7c6cc43156beba4d",
314
+ "b9285dbe600a0d883bb1a8106ab84908e1daec78d16497c41f6d64424cb98313",
315
+ "b97e8fd40a8dbf94cb8e2011dac90a9480e3f112d431fc70ab75ac357644f8e2",
316
+ "b9b5303b9443950e596764e3ab43ba388dc88bd9e9fa4db7feef060ce896c62f",
317
+ "b9c2abe9f6ecb2b7b77d64e7191d9ed09a36554e9421caf1b2e65c5463365230",
318
+ "b9fefd5cdff2e6c28088cc0a53e346a2c097998df8b6fc297222153e1daf65df",
319
+ "bb2bbefd2a016c0ad8e9e3e8858cd69d918f0eca2138b4f2289c7afd8ac1f4d6",
320
+ "bb30b5791e4307912aa9adbb690455c076e533c6285272da8d12e78f51db283c",
321
+ "bbf53aba9adeea04e3f958a95feecb13ac2d29b45a69478899314a996c78e544",
322
+ "bc6192bc3b2826b4e1b0b4d086536b6565c66d2c8bd66378de084b53625943dc",
323
+ "bce154fde165db7793c519eb7cd7b398ff52e9cd42578142e5c8ee8fb0656547",
324
+ "bd1dc7793ae70f7d999d0e13a56b492cda6f3ab0de6301699294ad15b001ac7b",
325
+ "beb32919cb900c53a3953442a5d071436449a4b9ec5b2bce6056f3d0a1f87714",
326
+ "bf6f5ac70503dd191546b89ffc869bc3e28efae51bf05ec330535f6e5286e301",
327
+ "bfc5863c1a54513bc8256de6b55beb115b2fa384a58f23530fcf523dcbb52e1d",
328
+ "c0b1165b54a3e58bae71e154a65337aed1918c1d065439d9a3f9d65f7d1324a3",
329
+ "c10afbedc26936296433deb12073cbd34317fe0ab7899c916c8cfe4001642ac8",
330
+ "c14e6d6cb9f0b03afcc8cc196c3a29e3e9e2306becbe004eb2313e306cd0a74c",
331
+ "c17f9ccd9fd2ea3d7d9230a42289690c9decb88e0a880a42037a79e94b100e02",
332
+ "c309e10ae79aa8821732639fddbc5af8d7480dcab2d0316dd331ad859a03eb98",
333
+ "c390b96eed7402670db4562842a2aaa9581848dee1538a66faef88134835c045",
334
+ "c39690ec883cace98ebcccef51fe48fc75b323082a2c0ad951b40531383a9717",
335
+ "c403e6f15abb1ca9d51a883ecc88cfcee37d64058e19559d87cc8e9659369ddc",
336
+ "c628c9f4b7dd0a2931ccd0a68cdb70335c49d125adc341c75ced526e0b5d7439",
337
+ "c66503618c5c0194e190365bfdd6f18fc85f6e8b14f248c0724575689225af85",
338
+ "c669ba0197f300ec35c9cda7a239e8a69e73488459a625ab9476b09624287715",
339
+ "c9ab26239629a397da5a94679f176211063e4fabfd3c41a2481405a5f70b2e71",
340
+ "cae0fe58a9afff0750883faf68e9d4ed511d4adf9f799828e6c7c331d0b2733d",
341
+ "cb0f5ddc19e8db59d7affe16f76ef2585cc9a4c992ec4aa144bd25fbaae42eb0",
342
+ "cb42cf2a13ecee4f5c3585d182f1a5e0172acde5cdb5dca20df348c61d5707e5",
343
+ "cb7117229c57ca9d821d8ec24d4a922d8348a5fe3325939bf84dfb18f11eb5a9",
344
+ "cbde24769f686efa9b3c354173679e351c5ea0d76fadb3e69593a7ead04a1ee7",
345
+ "cc70ac956671e95234876cc83614c2f88db9110ecce9381fd0ee64803b299429",
346
+ "cd36b417bafdabd7721c88c17d2c29c33e9a11cc0594f52a54950906d4a44fbb",
347
+ "cd47f108a726fc44a0eaaf3d4bde604dbdf8c41861f2c9a64ef11c7a36c8d6a5",
348
+ "cd4b3470071fd954fce44c84264d4bc9e61790b37d2b5e9374efef01316e6dfc",
349
+ "cda1f2569cdb637d0550db64ff59e80ba7330792ac95484c37cb0e3d113eccf8",
350
+ "cdf79c0142aaa8c5353ce2d676a01a9865903051dc987c8b73cdd73c76ec9c23",
351
+ "ce43ed1b892a3a7b12e0bcadb397a918718cd979e9e0795395c986f56bc2a750",
352
+ "ce7c329ebd2e1b1d9bbc17a9b9b74bcd914df56e03745413b6dce38eb4772bcf",
353
+ "cf2039c1d46672c9f9a5515c3d60575e6be00fcb93c2cbc1477459ff50955ef5",
354
+ "cfc4f4161ce90cb06c135ded875919f4675c107e786fb6adaa9070e9fe56d34b",
355
+ "d0ac396c83185c55d031ae1fbfa33bbb017d14a316cabb55503b6fe772fdee36",
356
+ "d11329eb550769a78aa2d9f8be77708dd9ed7527d706dad9fa2ed230fa38fe5a",
357
+ "d140b4148ccabf6f483642b7ce61825956b6e0a84ca2bd32a08a98f84a200a0a",
358
+ "d152d1675da0f33cf86b1664536034d4c9f4ddfc901e51c018b87280284756df",
359
+ "d2d52cebb51643d7f8ad3b2e52987b7c7f68e66e172418fb42636767f927ffb8",
360
+ "d42152e189ac72bdfc7e818137ba41cb5265a008aa3cdcacfdc560f5af5341e0",
361
+ "d49d739f974de58a57bc4f74207ae2bc8d8e66361adeaf55e961a27aa1eaae65",
362
+ "d563f93f87d17243f00fe477d8e78157ae7edf218e5c88e941cde003bea36a6f",
363
+ "d5a907cbc19ff057e4dafd6cbd075d983f7e71085148ddff5920871bc2167575",
364
+ "d5e7f6b0de3d88f2eef5cb8276e5f45caa1fb6e902025627b04b490b5d98ade1",
365
+ "d65210aa0d78bb32b64e066dcb64b6f8f71f71779a5d499354ddfeab22f1332d",
366
+ "d6f1ed8d805236ccf9f98fd65500bb4aa9ad8c41f8a071ddc29ca8e752929ae7",
367
+ "d7ad91e423e48f64697d40ac5d829b3386ed420ea0fd9841676704f0a062dbb0",
368
+ "d87661fbf955f5f80d07effe80080355aaeddb42c6cb0409ebf62ecb5393b674",
369
+ "d90b629f237efb49c3ab6d0aff93c838b4db299e6d7feee7636508d12b5c5fdf",
370
+ "d911da1cc1aa1a02712d5b1a762cf0b5d5ab389492c1a4b0559299f05493d3ec",
371
+ "d99afab46c17b26a07d8f71fedfadaca28767a8f350a2d9589736ce7da6cc3bc",
372
+ "d9df49c2c7600b22a4c93bb4b1f9dc9b1157f0117b215b7d1c00e26770ed5582",
373
+ "dab08cc34ed2fcdd470d03536ca642ff4624b381eaf9d536195bbe17ebe9c17b",
374
+ "db064204080f8555787bb6221183aa561add66a3383b4ba439f45c6ee5a2d96b",
375
+ "dccbc6279f3222dbde7759e12a7dda491c55472378491e24482e4ba527959a06",
376
+ "dd5269687a6c1f75d0234677f460ecef52a58ff78ab2f9ca6822a1316b36c5c4",
377
+ "dd87ae7fff40771e69e0a6180d1305fc6d1d0ce432e9aced610fb3f49194e7c6",
378
+ "def0de3ee663dc1c9d1b8df1f0127da4720d8dc16adffce81c060c330dda769b",
379
+ "df1f96b7a4242177ecb40fbebf6ad7091106e10dcb7b8be68cdf62ae9532d7ad",
380
+ "dfaa421efe0204df2a51dd0f34cc72c3310aab26426b107e3973b954cebb730c",
381
+ "dffea2882ee92b26abce07a5376e46308bc5e7a44749f72858eadc0aaf862fa3",
382
+ "e0077b61aca56e8aa2a018e68311f0e6aefb5d030952537cb7eef012f41369a1",
383
+ "e031d45f43b36e3034b2966c9f75218e710ec72dcc2a5e970c1d4244e4913e31",
384
+ "e0de0c7672d9da979ef118f4983eb9ad6ef0402fc8e6f5935c75326e26231f68",
385
+ "e1b16c59c51132e39f175f624b134609dc6280d818f8c47d9b6eea77e7b193bf",
386
+ "e2679dd5d22d82f7095519aa46615e702dcdf209163df9cdb734d295f70ed8cf",
387
+ "e2fa296a21818b123e3d7f367826c5e997b1371c858beb73610b110ca8ce4839",
388
+ "e50781f70c95585a91804012aa888669166226a18d9fd8d104cf901f565bb067",
389
+ "e579dd010621e89c430a10b2996eb3b216029e121a77237418f7280ad0f5d088",
390
+ "e61eb554944b8e39f620309e492c5ac7361b46cd9706619aa516e1c87ab2f11e",
391
+ "e6ad5e3ada47a44524ddd841faee42a564c8b8fff2eb255362771dd9f3cd5a55",
392
+ "e76923fb355d375ab50ff13eeb1ffa782eaae18ccabfa627716742e9c1390653",
393
+ "e7afca6e54dddeefdb0e7f1b02cc0cf2abf00561ecc8086fb882ed8ced4928f9",
394
+ "e84d29a0132a44780d83de149ff7581d54b4bc7acba7241efd3758dc613e3cc4",
395
+ "e906c0d38c20ca1f640313df9905bb9a39b48d7fe15c57a2ab75f1565a191fe5",
396
+ "e9173b067eb363df353546f8567b60b5122ced586ee8c343ccbb3bda6ab73592",
397
+ "e9795c286533eb785b99b4ec8b8d78f7a1eb8bb725e0d7916cafbeee1ca8003a",
398
+ "ea4b1dc02232d972aefbc800b4156284095112baf01168056ddc24b8a247f512",
399
+ "ea7664f177fe6c9fea15443fbf74a78eaed5618ae0f37bf189a0b726d17a5d4f",
400
+ "eb3537c28c26f4f61dee000695de989b41525c029c7d92eaff101a6c58fbbfdb",
401
+ "ebba25c528191213226e54e837bb727c0dd9fb48ecfcbfce91412f7fddee51a5",
402
+ "ebd7b1abe0d22ca80aa8f56e1d62c8a38ff410749cc212d3da44eb580195e0f9",
403
+ "ed4e3c8d856aaa25e7adb8ef0ab63276a38f2de8d211ddff05b794283c206470",
404
+ "ed82a193afed8645087d6ca4289eb20157008afd83272002a5669d25cfc51400",
405
+ "ee15e6d1acea63bf964c93eee42be5dd3029058f8770dd9caf4d5337a2c01bd9",
406
+ "ee5d303b876c4dac62384687b039f9658fff97c84fd87b08ff59b99a50a2c6cf",
407
+ "ee623702315a339ba49bed2fe17471ae117b38e2e55d2469e7cdd4c8ab88565a",
408
+ "eeea07e1f63be94a7ee57285c6c517fd3ffefdb1f48e02d1318b7d0d204db5f2",
409
+ "ef727e08aa8670e19394b9f923ac237625629c01368ef4a89e9cb565aad5d2ff",
410
+ "efc515168087302956545ad66a1b2923d107e4142ad40dc14c43ebfbc769e247",
411
+ "f08cf650a03124cb4ff3d44fd4a70abe0342eb048dca6644323b335770a9a083",
412
+ "f1176ced474ef2f0ade0eea28d45b69b1c172cb97124465766a696984397688a",
413
+ "f2143c8de49b21841348de9bd75cac7ffb13b71d118249abb1917698450f638b",
414
+ "f28ef034d38ffdec076a212d91fa453834eeab284db013d48370f45b23ac1e1e",
415
+ "f2f034448bd3ed4bf77478619b5c901fcb70f6a8f3ecab9dbbaa143a6953d9d0",
416
+ "f4d897d2739e313ae9a7f4fab2d18e96b751d08263efa31b92c65c39b569a099",
417
+ "f5636406724c3560b13585d91d6a1b58716ad37a776edc88af285d7b05f70d76",
418
+ "f56756762723fb4248c7c9cbc4288058fd1bf9a5b92dddc39c7454bdac885124",
419
+ "f582d63ab10b23f5177f0615badaa3359f0a0f7a56fb4165437c0ccb3bcceccb",
420
+ "f661f3a1841fceb3f780a7b40257af20ed3e704c463eeecf783b2c749d7a916a",
421
+ "f6685d129fc3f57ba35b71ce97af43735c428fee02c6c50fca3b8dfcce32da38",
422
+ "f72fe1265f2add1e7058bae000d66be9f4ec18762d6413e0bc58548f043a59ab",
423
+ "f731dbd87c85ffe281c1d8740aa242fa9a7b1d7210e0b1ca69592d11f6f11f8e",
424
+ "f7d9e96925c19e11d9a4329b0ff009a383d758857a44287115687ddc7d098d93",
425
+ "f7f300fb09d9d92cf906384f012ebe09835fb0d03b902e2a230c001b302b9274",
426
+ "f8fac33ac4e76d8146004ac03838c9e4977f31db2e3226693233e411ae4c5cdc",
427
+ "f919f6e81d3e9733aa6e62846129ac10370ca19d3e72ff8594876e9e77334efa",
428
+ "fa49239daee14b0d44eed2285a78ae6d7ea97346a5f71a9439f394afac870c63",
429
+ "fb3c26f9e5d25c1414b4eab1076ffe73023df0950ccb8798d2b2359e7060ff69",
430
+ "fb7d518eeb067ac77240f19ac6cece874b2af904d69198f83fab2f6f7de203f9",
431
+ "fb9f9805796702942ff8ce23f4b5f2e2abcecba678a82eaa95eb2f35d3d37cc6",
432
+ "fc3e326d60df78352c6da66ac491963b7c457797ebcd1e96c2436594fe59a6c7",
433
+ "fc6bb05f32006e045e36a3300044ffd7205edd7a2ace65ddb7fc77df71abd6f6",
434
+ "fc96267aab2baa5d3ae73bb0adc76accb567533bcbf6a69ddf358b7ece8e51de",
435
+ "fd03ea8635c19ccd8b6496bccce844e72697d6a638a7fe75234441e839b7615c",
436
+ "fdbb4ed5157a5396e2fb421e74d4d4cf4bf3131cca117026ea1f51d77a857994",
437
+ "fe4209d418bd8745b7e03e0450b65aa434d749e4f0b0f3793a1849ebbe1f93cb",
438
+ "fe54a9ab1415ed784e1c9d3aaa035570a6cf495a610171edc863f5743524e7f6",
439
+ "fe578fe58231dea95dce9a9201b55208564b4a59479d910d2144111d16e809eb",
440
+ "fe62a651f7a3b22cfc8af16729766ce8b0dba209d70e68d12dea3c7ff7c5d3c9",
441
+ "fe895f4166f609e50dcbadd349f7534f86356f5dc12f47d6511059ddd8fcb4d1",
442
+ "feaa87eec12904d90fed4e3ae7cd9189e9adf521af0c660ccfc811e39398c52a",
443
+ "fee98f221151bb980cdb862306da346bf3b0d9cec0b9cbdd9c2ba13ef635201a",
444
+ "ff1366cc917bc6dcdc630665d80888e067a67153e0291eb502ce1c061598ceb9"
445
+ ],
446
+ "include_outcomes": true,
447
+ "label_counts": {
448
+ "negative": 5134,
449
+ "neutral": 1325,
450
+ "positive": 245
451
+ },
452
+ "leakage_contract": {
453
+ "deployment_must_not_read_outcomes": true,
454
+ "nontrain_outcomes_are_evaluator_only": true,
455
+ "train_split_only_for_retrieval": true
456
+ },
457
+ "max_flat_action_dim": 112,
458
+ "num_groups_exported": 419,
459
+ "num_groups_scanned": 2873,
460
+ "num_groups_skipped_by_split": 2454,
461
+ "num_rows": 6704,
462
+ "object_embedding_content_hash": "411ee7d4c96458879f14fc9460c0e10d5309d83aacd5552c6bcbe089e81838bc",
463
+ "object_embedding_manifest": {
464
+ "dataset": "embeddings",
465
+ "dim": 64,
466
+ "extractor": "rgb_object_layout_v1",
467
+ "num_embeddings": 419,
468
+ "path": "object_embeddings_rgb_layout.npz",
469
+ "reads_outcomes": false
470
+ },
471
+ "observation_embedding_manifest": {
472
+ "dataset": "embeddings",
473
+ "dim": 32,
474
+ "extractor": "rgb_jpeg_stats_v1",
475
+ "num_embeddings": 419,
476
+ "path": "observation_embeddings_rgb_stats.npz",
477
+ "reads_outcomes": false
478
+ },
479
+ "retrieval_index_allowed": false,
480
+ "schema_version": 1,
481
+ "shard_content_hashes": {
482
+ "charts_00000.npz": "6f6ea9dd2f6809cbed119598842a1dd6baaf5db099587a2748a00264aafae4e8"
483
+ },
484
+ "shards": [
485
+ {
486
+ "num_rows": 6704,
487
+ "path": "charts_00000.npz",
488
+ "sha256": "538fac2468d1bde33b9e76f5233e331040a8c76ff5fd224727bd8140686f4071"
489
+ }
490
+ ],
491
+ "skip_counts": {},
492
+ "split": "val",
493
+ "split_fractions": {
494
+ "test": 0.15,
495
+ "train": 0.7,
496
+ "val": 0.15
497
+ },
498
+ "split_hash": "9d1dc1c10868f76fa080895df0f2730c173f12b6db14932599d55c2b60859381",
499
+ "split_policy": "stable-hash",
500
+ "split_seed": 0,
501
+ "state_hashes": [
502
+ "00febbd1a7ad126d88b65aac55a35a9c4c63bc11e0f6e45344ddd4969242b950",
503
+ "0161713e96d5c5e1a45422cbbd0367210454a00f7d6e9840771090f4849fcb5b",
504
+ "0271bb52312dbcc4b57060d87dd5b3a514bb13bb1e20b235ad8204d61ad50a24",
505
+ "0544a922f6c9c662f2c9238466411bd1942870f7f182a508fd124141ef777c01",
506
+ "054fa57f5f9a67e3839a4c9674f9146dda13d27aa43c99ab5f87dedb1c0c2712",
507
+ "069933a0983331f68db2455b6a2b4fa216a97e119e1fb5b42a65adc252aca505",
508
+ "06a95069882817d6afbcaae8ec6bc12ed361c0a9446c56639880a8b9223c0c74",
509
+ "06d2eae8a8bb4c4503348959d65191f1f652edf6267e750861de79a6e0f1b1be",
510
+ "08353e08c94f126c924237211ac6e55a78461a8ff88eb8a1dbddbf844b6f2ec0",
511
+ "0905ba70e353c374ec760960c11b7a7914e3a0f401bf94a7772225dc39a94cc7",
512
+ "0a48350b21b9cbc27742e61775d9a3cd611daa5905f96a2cc3c59625d4b081ed",
513
+ "0b7be6179dacba118f3fdbe0e890ffc2d4ae1d15a87a215f41574bda0b24907a",
514
+ "0c4be6df74c82811154373ce6212cb0d7766cfe37020a987e1c1a553fc82aba4",
515
+ "0c5278be095a7a37efb891610364a48e4a4d84d044f63fab14e15e9351d2a9d8",
516
+ "0cab5c0c4fd5dd2183a8bc5891151fb5e5697bb605a5f7aa3c85f87f0503a804",
517
+ "0cf52a33f7d5f568029fdd8a8a3b634cdc7443ec30e9c127a73f25543ccd2c30",
518
+ "0d7cc8eda8b418bc8993e1fdf3fd64353a35fafd2501660fe5d87c6f7ef3d692",
519
+ "0d873ad4e7f5c34b9002a03754159bc9f2641e80bb46292920d135468141f582",
520
+ "0d8db9d71f59cad390736c77c29a6ee9f9f24de0c5250e1b8115164f4bc33f66",
521
+ "0dc29c4c945813a6553cd08d2dddb3982abec790f14fca7f70dfab255302960e",
522
+ "0dcfcf107cd9dfedc537d2da4969654c2c4fc3d04b1176f8e1426633edc8fae3",
523
+ "0dff2cb77a11d1888d2a48a76c8da04fe4e0daa54b5194bdb5ea100e2eb48ab2",
524
+ "0e23f36f5e66b1e4a3a7c074a6757653fecb8ebf1beebcd1e78033f5c2656a35",
525
+ "0e3910690cb8e214f63534f41f2acc8afec60992e30a2ed4d2fb61fcd93604a2",
526
+ "0edee0a0681248fffae8e5677c3658aa6a8b59d9eb3e33d90ea3027923b85216",
527
+ "0f0d99b44408bc284f59809b1266764117c2426549ba6b7e22b897e4c24cdfd1",
528
+ "0f19062a067ca2c041986c3de45c4e8d4e0f32838adcf5b52c29dac097394b47",
529
+ "1023ebe09fd0537cc30774636aa5c98411ae6ed1c2dc259a2624bbd063eaf23e",
530
+ "1113eef231093ad70c840dc75559d4ca74bd3d8dc4aaefa1c0b04354fe6776af",
531
+ "11f96b224bfcd397ee4f98e388350e3b65418b00a9ab4d2746128b42b12b43d5",
532
+ "158c57e861f5dbf4fafa3370948c03172a7354fd34511a04af24d744cca2d11e",
533
+ "15a9a5593a8b8ab430be4f23e1279949867e92ff352f3e2033c69495d6f636f1",
534
+ "16a1c922cc898f22123974a540e044a4ab649945807f078770480f7297e8e90c",
535
+ "16ddcd38b720cafaa913dab508ceefddc7620486bb42e9e3b0a9b81e7155543f",
536
+ "16e7bf765913827062088fe3aab96fcdad44b4615afb788a2a07e434d272b389",
537
+ "171fce66778284ce4d146eac1ca6327cab37d27bbc4be27a2a9fff473038245b",
538
+ "17547f1d97d7d63690d4f5ae56430531c2bcb8c2289abdea53e0151d4bd9b79d",
539
+ "18d69dcc611026a63a4602aa482606f68eb3ce416d10ba04f9fadd6b3622a991",
540
+ "1921b5d9fe877551d561618aa96a78ddc9480216ba450d537d5e4b27206c4999",
541
+ "19cb279a5068d0754add0249c6c2577e07f80aa33c56746853c796f5418fc7d8",
542
+ "1a43f6367e6d71af41f099f5216e60ec305d43486fe887af0c2f19849f406b75",
543
+ "1a5d3ae36c55dc60075de4acb795d4258bd60953aaca4786039280f763703323",
544
+ "1ad053ad6219459720e464e61ba1165cbbe5210a308e3f80962cc0e4376e38fd",
545
+ "1c665a2c6b863f8415c790f358f9138f56958247ed05638712d6357c208708da",
546
+ "1da1e71bbc20c3fee4c7494196a98750f0081b55357d61ee7cdc53945bb7965e",
547
+ "1de189fddeff2424f521f5026da163632e4417fac1f9336255f0e62614ee5bb2",
548
+ "1e6a33aa31538f769c69f26afd8c3fb7c5f0290023735688823d00e5e34f22db",
549
+ "1ed8594ffd73dc1116a58995bf378e36dbf065e9af2b003e643eb9933e7422dc",
550
+ "1f7544e2474c3c34d37aa4fd972618a6f29b1e84a054ebe0dffc58c8a2e7a71b",
551
+ "1faf9206f8f302ab744b3d89effef3a2a055b0e600c874ec05045043e1dfbfa7",
552
+ "20191c4fcda51e9e83b7e86251c8a03607d12daf307afbc00d51d85e4e7def84",
553
+ "2053f765d67e1e99fa2f2d67b7c8dbdc0edda26a58e6602fd99f9153ec1e826c",
554
+ "207798bd670f5ca13a6fe29e77f180432de0fe8e09512311636a9c49b6d27523",
555
+ "20cc15006255a0f0abb3e7f4e59313cbaee4bb4c70b8fda8bf24a8429d7bfeff",
556
+ "213e926b1edf4eb61706e283cd09447dbb4e7e529420a94918b6dcc6593f0b29",
557
+ "2144d1c886bb5f05a02f9f782ab4ef7451d251b76cf8774d23862aaedf517c32",
558
+ "2171371427e17c1995dad1d3526082f7b27faa7ee8f8d115d3e1f5a29fe1bba6",
559
+ "21736847d7537ba2fbee1d21f9e7c6f55dd5c9df2239977d889770ddf12484b2",
560
+ "225fc26bfe12be03306bfd0e7e321879c3e0515185c5941d4162e432fe6ec2a3",
561
+ "22af0dbc179d12978f8a50de641f892441c0fa6ff71eca94e9a7d4fb6d805446",
562
+ "232da4dad7749216ed41234b7490922221df235896962a2155fcb14e975705d9",
563
+ "23466412690e3c55214d93f7681b36635a8e1359f0ede7506e7c8801e12a3fb7",
564
+ "23b377bbdaf00ea187f5864295d60d63f998e9923abd69ee7308b5224d0e37f7",
565
+ "23c0c827d7f0d4742515d5227c112062f3eea5b8a374fae7b9f04a34b8a1ddf8",
566
+ "25438255a311c51c782bd1ca4591dbe9229bc3be0160d5ef35b016acb0850b61",
567
+ "256f6415a7e248845e28ac5e7d542da9867ae0f5a8765a3eae7809291bec1e3a",
568
+ "279cef5e2ebb26ddcc87c670075598b35cbc4366b39e9fd22c16533fd9f6f79a",
569
+ "28a15d04c873a079c1d06c470d2d77b8e96e025c30b859f176dba529fdcc514e",
570
+ "28e0ea4b46d79646462b01eaa2c675b5872098bba0bf3d66a75d1c1a750261ce",
571
+ "290401966fd54d78073a1984e84efd046b150ad5b91413e6f5bbcfd35a27c19a",
572
+ "294c656a8b36af0bf9249985a8fdac65c4530e7c81d78a0b226f721ad6f9d2fe",
573
+ "2a018927dc86ad3fbecca897ae6ec63485d334434829d3670fe94776f8444bea",
574
+ "2aba3b30872a625774dc16ad32767e78fcc9e0a1e8e800f976967682c2fabf89",
575
+ "2b68052385906b5436a650c00b9c3ffe0253943c3a86a40187fb964d4fb2870c",
576
+ "2b7e6e58aaa56a83b488a80cf01f40bfd6fbd5941855bb29c2ce21221ef78374",
577
+ "2bc8e606bc7f0a2875ad9a588f3e7a0aa32df858fe5542ff8f7c102cc9c9f14a",
578
+ "2d0fab9e61b78cd62d1ca7d76326481635d04d61812c74ab9cbcc2b0e0a3ab1a",
579
+ "2d666e1b77c3039e2574d2c71967860a070f2166462f049a3245013d8bb9ec71",
580
+ "2d71b5cd66d133bb09dab6f0f35dd703104df4b77255a9ee1a8315d1ca6f59d7",
581
+ "2e247e7348b1c46bfa9b0068a40df1f45a4ab4d9c7ac366e14d0a9f99e7d63ab",
582
+ "2e7618499962df0d0c566b843a23f7f4fb31301bc9cbab802d427206f496cce8",
583
+ "2e89f06988ec2b92a1c8b18b2b3697ee0404bde77298adca7ce981dc1b1a7417",
584
+ "2edbff51b1020474d5e181f95bfd43c6fbfed555f6b5d5687917c0817d4ba0e5",
585
+ "2f08ab47ee8e608031042183ba73f76856f60fb0262a8adabbdd29700e91da9d",
586
+ "30928509ef611054e0fec9b07ce84529aa3afb560aab8a0afad741c493c3efc8",
587
+ "312adebbcbcb2b79ef60eb52a89d93e0dbced13464e3bf66ecfbec402be9c7af",
588
+ "32bb5ed54dbb72237062117185d38803a2821325ae9186773df5dbd2ee73155e",
589
+ "3301879024b9c79869a02253edec7b100d7677fcb2023acd6e9ae7c9cf46308f",
590
+ "333fc8cfc5f76fd98905846b73548b0023cac4f5f10cf49d3ea5efd8e0c15fff",
591
+ "3376c21147b72dc0474a3899c9eb74d175f9b78d228913dd7c3a27bc18f6af3b",
592
+ "34a2ee666ee8be188582c8f7c239b6ff96df48562fe53f34fbc9006c61b87b23",
593
+ "35f48e0a0b32a3966c94cbbb83fac82d9d6c89b6ae9194c446e085b265ef05b3",
594
+ "36178548038fbcdee704bc77af6a77590b37e6a3915480091ac1d77d9fe92a59",
595
+ "3622d0ee6c6d00a746bda15fb39d7044c31bd30e627a4a7ba5737477cc2c788f",
596
+ "3655a761324c8dc8844e8ba8b9a4772e2453ea55f7c1520925ef154dfd43df4a",
597
+ "36c71852489c6888128d3ff5c2e62d599759c400acb473bcc0a06b2f9acf066b",
598
+ "36f646f44b39a2775ef7f08cdd6256eb68b045ff9d2ef373e8826165c2fbe98d",
599
+ "36fa82aa1f5bee9ced8fb9c2bde5246529f5735b6aef2abfa3398721a222afcd",
600
+ "3718724484b8054be4ba240b10122a71068067812d05e7b9963932051445168b",
601
+ "371f27874971dd4d10050be4bbbc1e38e9afe1d8bdf5de5bbc6b3b480334e7ce",
602
+ "37487957af8670032f1ebd2698dffcfc99b5e2931094bbce10a1f5a41d17e5b6",
603
+ "376260250744115f58dd61526c1195fa32865622d2bf0daab219ef3491f00a8a",
604
+ "377d211e4bed8b5fe1de5384be5ba8ac62161b150784f0e7d2b6dc3568e4d508",
605
+ "37fdb61efd2db8d7ee146b537977800e6c42f93a35f1b452a8a76ae50aa2e1a5",
606
+ "3802165b910c46febd084599f3ad3bb68e7f94946aa316b3be4077dc64d6892c",
607
+ "380464d744f9af313aefefee8de18cddcbd32aa3e09aaf4ae13b60c0cdf2c535",
608
+ "385b3e8838fe0639b71bdb89d173cde82ce32e8218e8a6900f7d544855d3402c",
609
+ "3864afe56d299c4e93a7b4b270e10fb22a03024afefbbc4676456fb2bf09ee2a",
610
+ "386a215a2dedbc0166ce9e574929a13416acf7edf161c0c281a7d3752c1664f9",
611
+ "399ab40ccb7266269c21475263d0e7264b2b983c333db7aa09ca0c3ff532101e",
612
+ "39e3ae6e36f8d2befbcd1d844052de9ec49b87d041c1152f23ce51b5fe4a7365",
613
+ "39fc414f3d1216defb30c0764ca78e922de5fae956c8f6fcc0c6cd3ff3c03c5d",
614
+ "3a3c4ce053b1d397bbba192d6cfcf15b8c08f6e43f2286971a9321c4a1b9f783",
615
+ "3add4fdf06b87cfe5ea9fbb1f1a6fdb2b2f53181cb4f4819abaa511222c728a4",
616
+ "3b1c9478132c0351de3a4efd39668453b42aa1eb060e7f6a1def9379b74c4012",
617
+ "3b76ae93b0a9d50a284a8556fe33d6bd5255ea98d1c84822d166a4e5a2380a90",
618
+ "3c8e3b6a6980d85319a45619c74f75689a10d57455f60ad363b0db859fe33cb4",
619
+ "3e44df029d4120641875fb004a2a937c49fe4779a7ce5b1c15df1c7c1e3393a0",
620
+ "3e7dbcf1d6545bf86b069dd68aa26c7b2d4deaf965dca2f37a2c7b327ec007a8",
621
+ "3f27b12be9a2fd631200a9fbd83b29deecde1e760ec54035723fe0793de1e445",
622
+ "4136b9be1dde7532b54d1ce1f96598e153c15c3a76ec32b95813cc38b070b954",
623
+ "41498d0adc8e023aca47897192364c2ce5745b089f33013b8b9deb69e24b942a",
624
+ "4167b16627d6e167a2dbd1be8484490ec0c6da9de79b106634ad88b9b65e93b9",
625
+ "4171cf812123e2b6f87fb20b97a1992b79181a242809805cb68c7ec70913dd6e",
626
+ "41919df4cca85f0b531324cfffd6584be4830bc8bd9a988fdf7881d30a8dde56",
627
+ "4236092dcd909a41dc844bb5007d916aaf2142ba0c7332fbf96947b5b8ee61b1",
628
+ "426d33af3a135919ef5d2b06dabf99d1a1c15ed3b1bb2b9e4c2978a5a107b0cf",
629
+ "4372ac740d1727b2385f90171db7c748a03a6fa3849940c98f4b9754a7b29d43",
630
+ "442548c3a3f961d677d15b34974fe1729a5c922853972634cd4ef9a66a130b3e",
631
+ "446f95aee72b7659374dd227a9ffe3150e198e66df6434c3b4d8c9e563c35ffd",
632
+ "449d06ce2a53a15c8070a0555ccfe750c1f25fed1317c69aa040aa181a24353b",
633
+ "454c2f13e99ab69415a54b5e6b0d4d5f934892e1b5083e568be81894e1ca2a9b",
634
+ "464f92220100eb1bbd1b22e9d03626667f349d7ea6f74c23f1bcef7fc88d5ba0",
635
+ "47a9c4f96c6872687900bf48c444e916d883ae0147879a5136565d9f781aaf1e",
636
+ "4843e199e6ede40f1655b18465b9caba2d2d1afb569e977520703101f30088d7",
637
+ "49417479651902941b3e22068b8f19676e4f484938e9068df2d418739e62cc18",
638
+ "4a6cab077475699ab9b8df7a8e6207d267d0d2358da13cf6493010703c398c6c",
639
+ "4aa88d7007545b593c65410ddba0b240f986229a5b3eaf101d52da62d1d145f1",
640
+ "4ab7c6c04df7d5502cd7a523e550cff7e0e5cd589b392491a7331b95ee9da323",
641
+ "4b81a042ffa2df32f8f83b5604e4a00759eda66d89a404a2558eafba6d4fee6c",
642
+ "4bf18fcd7c9afb43f6ae51b4dc792f1c53af66240e23b0fccb23b7c3d8aebc1c",
643
+ "4c03b46384e7d52ff10e415873ecfa8c5c15e6a35dcad5767afed802c3d1200c",
644
+ "4c9e5cab65487fb267c99ba75713cfb3a2a030f49e6cfcfb4f49ba84f398714b",
645
+ "4d0131e19ef58b57be98408e41c9a58aa60c41ab4dd830f587cf209bd3972720",
646
+ "4d25e2d4328ed58c8f62e38711768a25b57fdc929210950785171937151217ba",
647
+ "4e5b04961304eba5fdde7a3060e1a8a98a4a72f1f260a71090f499a79769aa8b",
648
+ "4fbaa3eadac6794630ca6e6754e7f252387e419761876d548179007e6e70bdb5",
649
+ "5079ae092765e2fb5b517d123bd809195e2b2d26b55db1509aeaeefaac437113",
650
+ "507eb82aa5d9d9c694ec304b6dfc7e339efd292c0f36d8ad69de93022801fd97",
651
+ "50f8c0bafafcc8b5c0908d4a4ce21c87bf5b8dc7524cf3513aec538e4efde930",
652
+ "52bc6084f0ae006eb0c52083a34981493ce7918c8d02d4b5467b92a16cad52a9",
653
+ "52cf2ba1f84ccfbf1812ba17714ffb951d5a5503465df2fb7e14fd81930c82b9",
654
+ "53e0f8565233b4daba347cd65931b0d03d3be14b00b81f53eb19e8732f7fad75",
655
+ "54276b11b5dc801246bf871ed20acea86c35bc3e48d65668292c54894cda707d",
656
+ "5573e534e62463125b164ba4d53d7ff101f2040bc3e43ea9f8e66e8a78a66432",
657
+ "55c17cc62f79dcc544b49163b3e763a294a0b661482408b35ec48b2afc4eba4e",
658
+ "5797212ef801c231eeb5e6157caeb769c9f1c954c945c6e41f6b1de27fe6ea5a",
659
+ "58157eae91b262404e5660758e41c875ce3d2823b2e4e5fc30e8a3e1384561ae",
660
+ "5842b9664210226005965140e501df8cfd1193dc1def74032e3fb54e6b9217ef",
661
+ "590d891c164938e5e47d3dd069e1c81dced1f8a6e52a3020aa35ddce011f6dfb",
662
+ "599decd9be71c7b255562b1d9623c90819211305e579d8ad4a3af74c5ce8c4d6",
663
+ "59f04a0707f4139930d13192eb42ef9cf0a5193fa98f46e64ec927d3bd34f662",
664
+ "5a53a1e8f981c816e78f88933c5d2cf7aba4d9a9a28e953315d817abf3e937cf",
665
+ "5a8eec221fecdba8dfeba990ba2ef824a40da43c4167f70dac07ffd3335e8fbf",
666
+ "5b231a214ee5db29e6857bfe09c6a36d9dbd7dd2b85aa25528d47902099bd8c1",
667
+ "5c670f5657e3227b5ec3bc5199ed81e653deb046b4f6fb539c40281da5887418",
668
+ "5d8536a1f3f9d4054766e67598a6212a995f62d36f57fd8241d4432b800d429b",
669
+ "5d95de9ba01f727e8fad9c3476b8010bc5753d93b9f775ec94e307aeaab0488d",
670
+ "5dd35ce96ea763c051ec4c81442ea8b847f0d357a99fd385ca5b09e9cce36a22",
671
+ "5e1aee289d761d9dc1711bc6b3c97c0eeddc16b630c815839a156e42e7f39195",
672
+ "5e4eff8ba21f97afac6f8d2f6467b6e285e21dba4b21c6bc0f58d2f086edc2ab",
673
+ "60a111176ae35f70f9585afa440fe7f3d18977376c267480fa3094c3393adf62",
674
+ "60c84890cf7803f6ae595c3566b6aa4176405f2a8ba83c0f6cd6e000247ca271",
675
+ "60f630ba7d00adbbbe73b85429e665bd3c041464321cb346d27774365cab4682",
676
+ "61aa27300c1a31db1051074f03440d13ac491b0a5577f0b3de363e5a3b189bd9",
677
+ "6206882c83f707aa1f3259cd8cbc1f890f45bc4a85124cc602dc3318eeb6d3d8",
678
+ "635c92d494d7796b0c2f484147287f53a89a765e464d054cafb688b9d96655cc",
679
+ "6361a57a00da67d6102113b3f3ba0bbb7634c08cbd2096db4289f234380f6778",
680
+ "657d6b2557b1397c99b24a9e5777a3e28bb62b4a942ccf9da164c08e6a655e0d",
681
+ "65b892cbc5957bc2cc57375a544114a6475b8d9c274b5204d6e0d9c1b9cae15e",
682
+ "65de868c437f8dcddad0e37f6533940fde34b39148b9f63104f41d8145610798",
683
+ "65fc6b04f85a036c86076adfc04b741aedfe77a79a90db37b4023174ed4604da",
684
+ "6728ff8f1d45daa81b955edc59625766ed7b9e5b26f65edeb3dcd646675cdda1",
685
+ "69a10964b21da0fecd5c463c6816f544cb1efb299cfcd7a50a4b3a8ae59a9e0d",
686
+ "69b1edf50126e222eedc15ef65e9aef983fb9d340ccb13cb2740f509cb7019a9",
687
+ "69b58c84600892b4fa2cef6512cc93f04e14415058a2e2275bfdbb6ffece1495",
688
+ "6b71c1274c1625f9d19ef6d1f7044365d9dbff6bb252a6445648ff98c073baed",
689
+ "6ba83bb33eb0852b8a304b335ab26c5716a00bafeb0b1619146263c25ee0966c",
690
+ "6d4eb328e7293fa796088ce8a44db34d6e177e36d4bbb103f369859a08842c5b",
691
+ "6d9ecd81dd1ed4c5e6fc7bcec4e40d9097f9f818422e122794aaddfe260780eb",
692
+ "6dab1cd77a4ef3a9aee45a8537f58ab30d455d55700428a5ffab4022c2ee9d7d",
693
+ "6e29a3f512c9a33928ddfaeb68dae4c9ff946678744704d45ee053726b09e5e2",
694
+ "6e2d61579abd8f5acab73369a191a425e369300e3de717d6b5e8b1c79cfb2fae",
695
+ "6e6db234d66d1aab39325523447331e08d1255a7d81c67c10348b3abb03f24b3",
696
+ "6eea19a0e64ae3f4ec5dc0e5178341635ff858ec9c2f935976c054dafcdc05e4",
697
+ "6f76d0be63f1094f195d03f468ab82e36ae3f5f846d96d70812668b5d8e2f7db",
698
+ "6fcf3251ef3b9a95d9d2917e3604c3e2c1c05d2f4004fbe0686920abf3574575",
699
+ "71415a4b8c2ece02ddf3e9f1abe357c30c11029627c5510ec937582f36eb2939",
700
+ "71de47b052ee81350c5ad7224bc7dcbb5be7c4fd375f0821a479c31d82c9bdbc",
701
+ "727ca5e770a26fb5fa53155067cd05cc54bec3392d6a1856765f966ca660f9a4",
702
+ "72a6b851451517c6bdda780093ffb0e1f713db1580fb466d1dd390c2603f1d23",
703
+ "73e64ca24a747c2550b8a82dd413d5fe30e26f03fa95771d660c4b21561b7877",
704
+ "7498d1733ee3bfc7e7cdb11d6ad5a65369420da5ed5db9c4d752ae710f7e6e00",
705
+ "74e57ea458f6a044861dd23b053a31372df73e0b47e139b1c30d126edf775ee3",
706
+ "761134143d39da9361ac778266206266d961c8a76d60b063ff66903a0c51bb87",
707
+ "77e531ba2e0fdfb9d13b55ac7c023bec7acdbc08a87322fbec534cdc5020410a",
708
+ "77fe4e7237e2a146bc81019079776768d43ca9d55e19baf25f6db7ead3cb3fa7",
709
+ "7a56f9770759b68564c8bdbb38baa964b4acce93a3687488b18bf4a82561be32",
710
+ "7c419f28cac41a890964171a7ede35e4df22f5166d04b34ba2f1f5c530e895cb",
711
+ "7cb35abb83b346ffac1f0a58212082119a81f64f5cceedbd7b8af7810f2ebc07",
712
+ "7cfc2744e8e2c90dec0a973312a07f72b33ccff72595d6531f66c31ed099c6e1",
713
+ "7eb1f194992237cecaa642db7c5140b22a58ccf39c3aea4a21afb463b58b728f",
714
+ "7f60c45a4a714bf1066d53bf7d2b72abac5194fa5979d3aa68b93b64c9facecc",
715
+ "8243a06a68ff61d79c805f2c45bcd25fea70bea8fad91df198b432ab470ce9f8",
716
+ "8350ee6948c386f427a68e9cce71b828dda126548c1704c5aca0d040de3e1ec5",
717
+ "839775ebcdb6e011391fc3c3a66abd24907fcccd0d2c939b5cb390a65fb76f07",
718
+ "83d122ef72fc3ac252eefea7659e06a26f887d24374c04a5d750299f76aaed7c",
719
+ "843281e77ecc5a5210dbcd0b0b98edf7b95537fb6cdacbfb74fea2f8b11d8a51",
720
+ "85044cb38abf6ea983bce7170966ab3062fcdf01b28e25048481770a3418b82d",
721
+ "852cb696bf9bcb0a2c8c25bb5d111528a92006f3345eb8a47b3eef0efa343d72",
722
+ "85654a7b2d601c501d06ccec3bdcba969240c6862c59a34fa4ee9e3216a6f978",
723
+ "85797e1a9396533183d3551f5f62e960350b2cc8692881077242461722204027",
724
+ "86f1519103b9098184c2719a2553fcb4f24abf9635e7db4f4e09808b8af84f2c",
725
+ "86fde29349d8159afd645bd5b7c5556bf74a7563f5d27f5252cce6523681f6c8",
726
+ "87100034cff240e23384afe44efb0b09e90603f382896bd58c186faf0fe4aeb5",
727
+ "8751179f0094283c5cf1ddede063340f3a84784459a23264adc86421baf930a9",
728
+ "87d27f35acbb6f4734a61bf3a13edbe579fbb5d4ccf246ab2378e24b7558253b",
729
+ "88b7ccfb5fa2d3d4bec816f96e0a2dc5a15ce4ea602df297725a58dcb0806732",
730
+ "896df52c27a5c01769e1e6e9fbfbb795f43e7de78ac34f34a40bb37c147de197",
731
+ "896e19fbf4b3372093f28200f82ff64bedf1d38b1f7fe349d998496342162082",
732
+ "89e41d06c9b85a5e85183168ae099df4f324852eed8ec1314e347767563fea96",
733
+ "8a69885fdc86627e9cf7ec2880a8c8348b0cd02df31962c0c4d2c389a1a066df",
734
+ "8a8bdd653602f2f158019a5f439e3a899cf0a85921aa14cca082d6e603560909",
735
+ "8ae4bf13a85c96359e899a85b465784a6bdc84f8f36315f706c34d955beabd29",
736
+ "8b6ae722b7977d64014fa0af8f179df23ff208fd2bfae650d6c2277dde6979a3",
737
+ "8bc2f96c2f5cfec2923d02f631b6adb72273f0596f19d1de7b39d20acf0b6b12",
738
+ "8bc5b83371b1c90418b00374bfd06f8db036885a988bd6833ab0c06fb1ea98df",
739
+ "8bca1c6617cbbfd5ce3cfa106987cf26130f8539c26044e1a22a2f51fb1aca1f",
740
+ "8c05750c1ca939baf32d25d1f2d92a61800b49a76426ec28c1559d59606a61eb",
741
+ "8cb8ad276815c315802c4d97c5820cee5b1077c684875ebebd549432c2c4a812",
742
+ "8d091dc0975a50221f3b19a0569128ca17b1a52d01fe047e380b0aaf0e63b52d",
743
+ "8d314a3c1bf7b6c300dfecf79320d58bf6d2c23fb174f92d5c64d5e8542a96d7",
744
+ "8fd961b8dda83744a40dc807e6d7d2a3ea9d5d1bd799c1e1d5b2fefa666d6ec4",
745
+ "8fe2c0bd9c12919205e8c8b29ea61bd96d89616e9b7498b2ee8013446d5dd10f",
746
+ "902b91e3a5080e9f94da80889ceaa51c5f31d964846008622b6ddbeb07e28392",
747
+ "90e57c604e72df20a9694ebcd97e3fd9f00c5e5d8ec239e84a7b8a0c431e66f8",
748
+ "92928d4b346d02eac80d60c98e94f9f1b113f7d01134383db932754c4c5bd3d1",
749
+ "9393b510d59ccd16774e17789b3a01f90a64554dffdb9eac197d050c0bf5d09a",
750
+ "954890c63eef6f575f68e4f37adcaf21f4a2e343bb291d0b5585d0ea6d46a6b3",
751
+ "954c283eebad88edcd070cdb36ab55e457424f3fcb9210f59386cd3e4ee4e74f",
752
+ "956fbfd3ce94d4619ffdd9194deb85dc6d82e89fa199cc978e0cfca1d3e8ad6b",
753
+ "9588b1651ce73d10cae6b77ac6b991868c806a63979dc3b4be971f686e9723b8",
754
+ "95af0aec0162909a0addf5f0002455abb2bd61577e28562a5ed00c2260cb8fbe",
755
+ "964083a52374c2723dcb95bfed5636587a298497bf58e1183b60af055c46f397",
756
+ "96909c33a461e46a0d938d888e4371d1295f04535166bc16c9dfbb35c9d552da",
757
+ "9bd4c58f5a4d057257cae54e018f4eb88f9816f3cf7f88949345453a70512ff9",
758
+ "9c674f43e5ec1a37fb7eeb8264a5b4d785f17f716f1897c41616ed37333ba70a",
759
+ "9c6faadf417db194a1a2e68cfc6a4792339ca759edfe44c747fb4c516d7a66c2",
760
+ "9ca9acceac4275ff088b9bcd89d78019a7f71f284365c0681882194a8af11390",
761
+ "9ceaf7a3ba12f73c742a6f86a81c708c03240515b9cc5b89cae86ec4601a744e",
762
+ "9d8ede4ea688ed3e8334867c8f34e34e5c070ac993ae9bcf04489e7a25aec07c",
763
+ "9dee4ee61296b82b98e8972d47e802ee8ac8a5008b5e03cd5b130d0edfa15c77",
764
+ "9ef46448570d1e27b98284d64a08cec2d25a8118fad8017d44a99f4f6c061c47",
765
+ "9f1bd954340e3d26d52465ccbcd4b72baa30a3ce520e2a6e4358b5617b176963",
766
+ "9f585ccfb9a92e299ff0f76a9a777b4c5eeec4189f4fa9ab50c1cf3d74c21e05",
767
+ "a17f4334476d179113e2db7a6dbff71ecae86b10dfb2ad7fe2d8cc9135b4f2a5",
768
+ "a3170ffc83f7e0cc197350a8d74d85500d474673998ef1c62f1ddc3f364eb53e",
769
+ "a31eca43deeec05441c35933fc20b48f45b191042da018d7f6d97232c1ed9ce3",
770
+ "a322d6d1f0b71546ad4e9f746bc3fbf1742df0ecc431dca06584e44f29328e43",
771
+ "a35f0920ead6068375bb19cceff473cc971e9b578a0bc963fe7e617e2fdea9d3",
772
+ "a43f30a7fdf9c7a1561c6eb9aaca66890cdec58295265ab8d305ff076e36e0a8",
773
+ "a467e1194c7dc803eec26bc9535be71b58444bb5748d179e2a4e5e0662d3e0de",
774
+ "a49644cf53180202b00609f3cddc005a01e2086bb63369c711ff0950c8f8c214",
775
+ "a4b7c79b9897333ce6ceb761dc6854a40225ee9211a54101ba64dcbc92e0a6fe",
776
+ "a52945e1d0442f75f9a004ecb76d0e2b17e6137b124e73953d23c99aefa7749b",
777
+ "a52d52c4694251b5798e3478578d4d39fe1e964a70548f6388cd3c5e5cdb0fa7",
778
+ "a53fbb7223f7fdf1f7b641c51f0f8005ec0f32cfce74d9276001d8d8ea1151dd",
779
+ "a59577e4bbf52734f1931725d54881da6d66231bc9fe4c238d840b8c5444fa4b",
780
+ "a59ba23421ebc70cc23935fec64df1f948ef7e38c9ae6ae5d51491aea195046b",
781
+ "a939547a17e2245da625a539382110eaadcb4aa5b5e0821c15d6ae3a017f55e4",
782
+ "aa1da523736c9af8a5c7af126aef2900c7e06c0ba501113ecaed22f8982f7cda",
783
+ "aac283530a6c5a5e222a951421b18af4933a1f893ddc009d2e500340dd5d5abc",
784
+ "ab4a571a4eae6c30751bdee0aed400dc75363f94dc0f4b7782cb4556e83c23f3",
785
+ "aca0c3bd9fdcf931da31abcc3e4ec560120c83ab86deee1e627b5fe8e06b21b6",
786
+ "aef1e30af94b53299662a893e7d3a08099e1a9e2431e39096bd55d3eefc3953c",
787
+ "af28f08a766c0dfa0650194c27fbfb2b9b057e676e959dcde8c0dc04ba0372a6",
788
+ "af680a60c99d6ed6f75252add7208bbac1f27c4263be29008866b314963f9933",
789
+ "af8587819a0c45ae5a25878cc8bedc571a5a14d719ff7ad25bd3b088af294b3b",
790
+ "afb57d77ba6a109f6435059dac6617930c19567b3a9a5c690932e04ef7b7d45a",
791
+ "aff6e17d88bd7ef94395eaa52ed7f8bfa1a84a322d28c1535cdbaeff3fe137c8",
792
+ "b015523bce4598e2b2d077d075a5eed6fae4235b2851862ec3edf2a2a454b89b",
793
+ "b018d81f318adafd8aeaf9997b18e1e23da168b5fc28e640cf50d570dcc9e315",
794
+ "b10f856b83f430458887304c2f3609fc5bf70132d9546fbbf763183d9c8ec5b6",
795
+ "b12d970acf0fb01d206f2547177e9c4008abb0d81d20a8df2faf566176cb042f",
796
+ "b1daedac8c8e33c5708b2cb66bd406e57633cf05b6bbd95c11f9e752b35da4b5",
797
+ "b246439be3988b28f0fee9316443e3a92a37bf97ce2be5c748f04ddf800418ff",
798
+ "b3ae5672ebbcf36d12467c9fefcd760018e4b5733a34c41bce09e760bfaa9715",
799
+ "b50ce258f9c2c3a1b1284b7a573a2fa4a11ebb787929d39011a47002463ea7df",
800
+ "b50fca44807c252ef01dd95698262033154f8707905f779c0a20abc6439ca0d9",
801
+ "b5a7e7146a98035523ef6490183cdde177444103df52e1217bf8e72ada4d6328",
802
+ "b6324c2d34799565940dcccdd23e1360e974a46b44c209a6a7a9fdabf790d044",
803
+ "b66654f95bae5b30f1a326c90dc930b954368834385eaa21725571afe94b354d",
804
+ "b677f0aee329cf1b061cc0e2601a9c33218c5c10c0cace5820236313a0825d3e",
805
+ "b679de5b8df5d6b74ce1c8331418e05575117dfb8903a5381b6e92d9eb0b6a4a",
806
+ "b68d92aa8fccb4f75dc0ad09e9b6d567f7d40d421ae5321b36bb0a48236e598e",
807
+ "b72fa8f582c7fcc689962e8ffd5f8fe3f837aba4cd11ce7687532498c11fede9",
808
+ "b7761c8e81ab553d5b43ef3e570e6639d8d4fd3d3b217198f7bcc86b0078db4d",
809
+ "b7f2c74cc6970aa943740ba5c7fa947e4dc991233fd35e51412ced790d20298d",
810
+ "b8c14ae75cbc0d0dcb2d256815e6e4fab110ea29a3582201c0df5cebe47544d5",
811
+ "b8c87db8cf709cb8f4bbd8e684279f796def8980d809fd174a79690113de7207",
812
+ "bac19e11a6db7315a718dcf8236ca104336de292ca5553afe82f9418239b4271",
813
+ "bae7c6cd0a3c6e5f44834d64879b260dece1a4232f3fcb1f01756b03391c8831",
814
+ "bb5d7e079e9ab21431f60dd6c4495a3984001039620203c0962965e5f42de079",
815
+ "bbd5c9fa1463d0d789c66b421b274e424b1289de879bc2095a39924521dcdf06",
816
+ "bc107f5bfc824a5872fb29eb153fe68133664f84a0bda0c1f040629cc5722e55",
817
+ "bc28efd0ad227d98655ebf16a86d39e0eeca17d2f7e4a1bb14562b9772777332",
818
+ "bd570017cc62932d88f56095bb4fc99aa0d3f718317c089808b4649309f57322",
819
+ "bed8189cf521818dded6793b749c12a91df30fa033b1d0d628b74b078aa3d17f",
820
+ "bef655041d427966b37d503ec3ee6a608a397e34e50b3faf19c7c8010d32e9e9",
821
+ "c0b19497f20c96ff5a998cdfd8fe901b2636548e67f7521e17d8c1c3a344b119",
822
+ "c1fe0dcfc510b321f89d86e9e28d959be4b9f7191bbdfe75d94aa00ba907fecf",
823
+ "c29fa01c7d7c2300dc9f248122d101f2ae7096dbdff573efbb51b271db18a237",
824
+ "c2c4d1a16f5c13da6ead578d5ea39e79319c2fbe4b950d9a9eef447f0d885985",
825
+ "c3207049b4fe55b2130b223f9d327b21f11dc51c5ca20eca2a8d75c13f6000e5",
826
+ "c4c4cd450d60c5748c8fc36d10c75f182cfe952711617dacb76a0b0e7391a35b",
827
+ "c53776402fe96a74a9aa02b2ae732f279a3ed258aebd35e7f7347da28491856d",
828
+ "c5608d269923a2a3ee69d360c52f71f1eea591a603304c908ddb465f737c0782",
829
+ "c56cf2d189daaefc764ff1bbcf667d4625be73dbcdfe9c42e0f503d59b5ccd23",
830
+ "c60a678a37f48d744f7ba785c4adc7de65317982199801ebd603090315c85ca6",
831
+ "c6244a7b5686ad0e553cd0fb349481c14e4390ec7a65d414da7dead452a535d2",
832
+ "c6dba1af7f4d12fe2720b67f0ba6ce4984cf9d47c6d288f2467681d02f5af6d4",
833
+ "c7418675ff86e0833d60ea80a8c8a0f9b99d239421ca68164c6824a99a276f54",
834
+ "c77af3500dc3ee06cb5dd5ef1e7cffd70a3c7bca08b9c00196884cc3dae8db90",
835
+ "c9feb7939ea3febccdec91c170527060563ee0dbf0dac5b75092ef5c5bcc9e6f",
836
+ "ca67eca42b424c6374979d63df4eeae00381446c7c7eea0934a3a21b9352cd9c",
837
+ "ca83a160910b077d3ca4b6f0185b4a99549d8287d187a4c204ffcb6867514887",
838
+ "cbddadeb108dc36a2cf5e686e2cb3a143946afe47ceb15dcc185af4b272e3777",
839
+ "cc47d275990836af665da5f4dcb0c118d3d4897d376ebeac53ba6f207dbb8a64",
840
+ "cd015dbf90e521f2443eb714d30785c1d21240e44dc10e1f8464ff146866f6d1",
841
+ "cd7d0bcd5fbfa262877644a5d88d8e2b6016e6cb9a880761c1cc1823c4e6c621",
842
+ "cdd8100bc160c8c866913e8a033cce5f402efb1fc1538247d4cbcda3e1216659",
843
+ "cdef8f05ddad0db32cc966def25db6898a05c42bd15391b77c619eaf5ab2a904",
844
+ "cf4dd171afff5c96a512d37c25ea9e292f1cca25a2d267b2aef735d51605de17",
845
+ "cf99daf73630fb1d3448b7bbf0b4ba0558c886dd8ae1fd77a38a0b610fa1ea3a",
846
+ "d0a9e8772c316ac795de39d315b5ce38b72472b4df16d6ea6ae473c94c853a0b",
847
+ "d0c4b2351028946098f41c3fe28981d2aee8e3f5caf4ba4fcb12462d0c2fa12c",
848
+ "d0d11e3470945260027f6fefa6a4d0a3cbd29d2ecb9b9420a14ff6fa952c957d",
849
+ "d0f6f9c2fb93b869433b409b0093f0e15e226e6d750632a90a0e6e970a3464b6",
850
+ "d1ac7a2471d394057a588ec47c74a0b78e7cc2f495bb1585b184eb72dfd4ccce",
851
+ "d2615c76b309b1437f0d1f0294b8e88fc5fe3e9259a41a9ae4cdf137ef73609f",
852
+ "d3723031d231e59cd40ba8632e01ec7604f5197a48a45898cd0e314541331c87",
853
+ "d440f8f0db5b04f2ada14de102a3b3a7e3e774ab487a6bde0275a4344d62dd16",
854
+ "d584d77a56c70ad17c27a192efd074d4f5f74452f2c62ebae6b627d80ae063f6",
855
+ "d5b1bf93014f200928b481640c3ae57597cf73a0f1668a77e07ddd8683a81508",
856
+ "d652ef33b9ffeeff77ae17aeeea979ab8c1d403431bdf245867abf4566b54689",
857
+ "d66bff5b50f3ca994a03faa77246bf9e3a8193a4903fec25d5f56f7fee31611a",
858
+ "d66e373e83c0d5bf391ef83eb47b2ea4971409df16bf2d2d3f90e4819c641cc6",
859
+ "d723c5e330147c034be5d384e3c539f323c50b01326b901b3f78ec8a29d13b41",
860
+ "d842c9b772dbd601bfad1c3d830df5aacbb685507247666a80d8ef7ec0ab4db2",
861
+ "d92d42c9d643429671fe1f7384014ca15ae89ce49be0c5d524d2172938585986",
862
+ "d9efd2aa3c0a57f3bb93916653cfee963e4b1141917d703e61509d8d6dab7c78",
863
+ "da5e94dc767915969e1c8188d29035197a22ba0b9d8be1cebc04e5a8203646a0",
864
+ "da96925490c07aa86ee90d9b7c31bd59fdc56d27c891cb6728059322add4260b",
865
+ "db258468bee6ca4d5ca59334c4e45ca958aa6df89ceadebead6c04ec77feca35",
866
+ "dd2f29960e09d45097d7c7324130b96c27ee8483c4dca1b0045dd58a158b4838",
867
+ "dd794a7a3fcf1fe4a062e9664df3eecc6f45f0f15bb1672fe552f117acc1f646",
868
+ "dd97f6d3742257341118baa40d95ac216a80d20d39d47ec9c17e37234151d783",
869
+ "ddcb6f80de7e1ee656dabcf2566d3a78fb97806bbb6077a66d0b1e9d8a3e251c",
870
+ "de0beb18c2c3a2c895ca5f7a01665041577ef8373f8037591a1b1bd9b906e13a",
871
+ "dedc28fe961870e02089cb419fe3be918a27dd6568af8724f7e6879f749cd029",
872
+ "df67a5c802fc30ab72aeababbb0e93dd8e5b42cd9646ebc54ff9f47240c6f361",
873
+ "dfaa69ca87d3288942dcaf896c9ba8b8a6b0ba61006631efac7e521e8b543c9f",
874
+ "e128aa827a20a9ca1a057cf1991d49872999d78244f9f9d126399e8aa70b4841",
875
+ "e170828ddffedb42c27cf62dc347fca5ebcde643e3b4585691189c321cfda9c4",
876
+ "e1bb15a4b9aa87ea3327bde96322b226d9e407fee96824e7281a4142d9c32592",
877
+ "e1c5715b0a722ae7aa244fc75ea30dcfa239f6af640e6ba5d4937d22c85cb982",
878
+ "e24cb331e947114a105313fec3abee7bba02e65792962b11f789ff65354bca55",
879
+ "e34faf75bfbb6ec5d6fea460d34a300f33d61d5544249ce0adb97e3c9a8bf469",
880
+ "e417c01c804a4b846fb4ab0bc5f5e635822dd469814b6f454344a371a0c54846",
881
+ "e4a3df05b8fa77155ebe3f0e59b48ed9727f70e8624623cb69596dd0ccb2acff",
882
+ "e55e578cf3630b4265a38473546efd00faeaccae3b69cafbd46c38b43810159b",
883
+ "e58a39b6c5e754b1bfd881b5e6b3072eb678e7909a173dbe589b1f1046cc3c61",
884
+ "e614a0919bca602cb5416a1b69b025e7679766f28a630f8fe36a044c9b6c19a2",
885
+ "e730a01ae3f75be9423d32d2fc782f0250689840a9f35f552f9234956b91b651",
886
+ "e7cbfefa2b400f6c2a4088c06e774c341f1fec283d7e6de39d6aa38f8fb16eff",
887
+ "e83f6805a42db291bb19df80219ebf6257f4ca9965f7c96acf9c84c19845466f",
888
+ "e8b62f0abf8484fd663882a6423adc003b27b151b2e081544cdb328480b0efec",
889
+ "e968c0a922d5fbb0063602dfac63455d45924717d817299133a56a73fa9c85d9",
890
+ "eb68932633b2c16e1761b398a52354406182ae85cbc65ae7e27dbcfdb667d6af",
891
+ "eb961efc66c3dda1fceeffc0f5dcb270b6f354484fe5ff983e7252533d2db5b6",
892
+ "ee27e25f902e654569ef7a8ded0e028eea1f7b48cce472d02537f9c58b90f31b",
893
+ "ee9c47b6572c11e77808aa5278255d5baaa53327776a2eb1f3c02ac21706667e",
894
+ "eec2faad73968e9d8df0cc9171b24b238365533e490c3c1da68f2a01b3a58dac",
895
+ "f01042ebfb2d85430a7264450902c9d7f8f208343cbfa398de5efe062158cb10",
896
+ "f0262b9ee515bf10133044931f0078e678627bee08a7021dae6532b1d10652eb",
897
+ "f07a5ec90031146f8a1b92c877a9ef88d90b67725b687daa47850afcb8a4f5bf",
898
+ "f137e08187e31a78bf93686f27df66e24ba62fb370115cf099ab3452b3e84934",
899
+ "f18c1737e19dbb670db3262115d9d640533cf10ba73f8196ddb5a7d582744feb",
900
+ "f194576d56014f50ba3f94c92baf120c3dbcebe176b661154484c2a92b7b40c5",
901
+ "f3088bbb09667126643e781a17542b06cf97833d29b414ff29cdfc0c249b0025",
902
+ "f36b21cc9f9651ab59717c67501c154929fdb69d2124fcf87beb5cb4eddc385b",
903
+ "f37badc006085509e7f393cc62d362f41a6300fe2a54c46c6e0763b4afbfe7f7",
904
+ "f3bc1c4b3902b647d6660ae83b6a272dbf80b85cbcb7b57d74afbe94d6d754b1",
905
+ "f54e7d6b2f7b5aaa89a466187ae089784a605b946d50d221e21185b14ad2e5a0",
906
+ "f5579f35655c4c99b49931e63782e64494470ac21c1693fecacad3d3d54291bb",
907
+ "f70bac72a9d4055c60922c63e2f2cb3b77aa90250cc071a96f67d3ab658a4a12",
908
+ "f7899c890521825ae7f31b8c107f2ac9cf8e50097d4c06b92e226a772ba7b925",
909
+ "f82052e29e6ca52c7c633334e5c939674ca671cf5a592db59f882e88c5bbfdac",
910
+ "f937b82c14d253a7eecc1fc8effc545517af158c38d181727dcb8205e42a27bf",
911
+ "f9576ef7044564daed5e3bd01e196df3a69109244971a106c44d8108f286f08d",
912
+ "fa65af636e0484758f08f422fe84c5ccc340fd7f33a96fe5bd63e2a9c110fe6c",
913
+ "fc969058a53f8827440a2d1750091bc41aaded65e487bc033c81f2140b997321",
914
+ "fd18f20b9ea23bcb6fc5091c74383ba65a8b58c19f12da01c74b82ce5f821843",
915
+ "fd5336c29326b4bcacb4c55cedbe14a1d5e61ef269102110f643f665d1f8f03e",
916
+ "fdb70a2a5773acfdd00811abbadb690df8422aaaed7bd93893f576e2f5704bac",
917
+ "fe2303f90cb10eb75f13d551f88d2103bfb4f0c83682f807942f4a768b6852d1",
918
+ "fe6e280d14d7dd0fe23a487c684330c060896eee257e80121bc188a4f6110c09",
919
+ "ff5cbd0b1fdca636b746e9d39e0c2e984969c249bd093bcdd886290622f87775",
920
+ "ff6bbc3315c35a873ae49a7aea0911da2039eb0bdfd754beec5b119d5ca46b29"
921
+ ],
922
+ "task_counts": {
923
+ "LiftPegUpright-v1": 1056,
924
+ "PickCube-v1": 2384,
925
+ "PullCube-v1": 928,
926
+ "PushCube-v1": 1008,
927
+ "StackCube-v1": 1328
928
+ }
929
+ }
workspace/runs/ctt_residual_base_context_obs_rollout_val69_perdim_trainmax_seed2/train.log ADDED
@@ -0,0 +1 @@
 
 
1
+ CTT checkpoint trained separately; rollout used checkpoint=runs/ctt_residual_base_context_obs_seed2/model.pt