Subcritical Percolation Jordan Readout / bench_report.json
Mechanism confirmed, baseline not beaten
1{
2 "bench_version": 1,
3 "track": "custom:csbm_radius_graph",
4 "model": "GraphNet",
5 "metric_direction": "lower is better",
6 "n_seeds": 8,
7 "baseline": {
8 "best_cfg": {
9 "lr": 0.003,
10 "gamma": 0.0
11 },
12 "sweep": [
13 {
14 "cfg": {
15 "lr": 0.001,
16 "gamma": 0.0
17 },
18 "mean": 0.11812499724328518
19 },
20 {
21 "cfg": {
22 "lr": 0.001,
23 "gamma": 0.5
24 },
25 "mean": 0.11812499724328518
26 },
27 {
28 "cfg": {
29 "lr": 0.001,
30 "gamma": 0.8
31 },
32 "mean": 0.11812499724328518
33 },
34 {
35 "cfg": {
36 "lr": 0.003,
37 "gamma": 0.0
38 },
39 "mean": 0.11499999649822712
40 },
41 {
42 "cfg": {
43 "lr": 0.003,
44 "gamma": 0.5
45 },
46 "mean": 0.11499999649822712
47 },
48 {
49 "cfg": {
50 "lr": 0.003,
51 "gamma": 0.8
52 },
53 "mean": 0.11499999649822712
54 },
55 {
56 "cfg": {
57 "lr": 0.01,
58 "gamma": 0.0
59 },
60 "mean": 0.11562499590218067
61 },
62 {
63 "cfg": {
64 "lr": 0.01,
65 "gamma": 0.5
66 },
67 "mean": 0.11562499590218067
68 },
69 {
70 "cfg": {
71 "lr": 0.01,
72 "gamma": 0.8
73 },
74 "mean": 0.11562499590218067
75 }
76 ],
77 "full": {
78 "mean": 0.11843749694526196,
79 "std": 0.01745250183842921,
80 "per_seed": [
81 0.09999999403953552,
82 0.0949999988079071,
83 0.14249999821186066,
84 0.1224999949336052,
85 0.11249999701976776,
86 0.14749999344348907,
87 0.11749999970197678,
88 0.10999999940395355
89 ],
90 "n": 8
91 }
92 },
93 "idea": {
94 "mean": 0.10937499720603228,
95 "std": 0.01184205993180543,
96 "per_seed": [
97 0.10499999672174454,
98 0.11249999701976776,
99 0.11999999731779099,
100 0.10499999672174454,
101 0.08249999582767487,
102 0.11749999970197678,
103 0.1224999949336052,
104 0.10999999940395355
105 ],
106 "n": 8
107 },
108 "comparison": {
109 "delta_mean": -0.009062499739229679,
110 "idea_wins": 4,
111 "n_pairs": 8,
112 "per_seed_diffs": [
113 0.005000002682209015,
114 0.017499998211860657,
115 -0.02250000089406967,
116 -0.017499998211860657,
117 -0.030000001192092896,
118 -0.0299999937415123,
119 0.004999995231628418,
120 0.0
121 ],
122 "p_value": 0.2165,
123 "mde": 0.01519513610205344,
124 "mde_rel_pct": 12.829666696752422,
125 "verdict": "no significant win",
126 "system_worked": false
127 },
128 "mechanism_signature": {
129 "predicted": "Jordan candidates are selected consistently across independent percolated views",
130 "observed_candidate_count": 2.0,
131 "observed_model_system": "trained JordanNet evaluated on the registered root-classification task",
132 "confirmed": true
133 },
134 "idea_grid": [
135 {
136 "cfg": {
137 "lr": 0.003,
138 "gamma": 0.5
139 },
140 "result": {
141 "mean": 0.10937499720603228,
142 "std": 0.01184205993180543,
143 "per_seed": [
144 0.10499999672174454,
145 0.11249999701976776,
146 0.11999999731779099,
147 0.10499999672174454,
148 0.08249999582767487,
149 0.11749999970197678,
150 0.1224999949336052,
151 0.10999999940395355
152 ],
153 "n": 8
154 }
155 },
156 {
157 "cfg": {
158 "lr": 0.001,
159 "gamma": 0.5
160 },
161 "result": {
162 "mean": 0.10937499720603228,
163 "std": 0.011907849530802815,
164 "per_seed": [
165 0.10249999910593033,
166 0.11999999731779099,
167 0.11749999970197678,
168 0.10999999940395355,
169 0.08249999582767487,
170 0.10749999433755875,
171 0.1224999949336052,
172 0.11249999701976776
173 ],
174 "n": 8
175 }
176 },
177 {
178 "cfg": {
179 "lr": 0.01,
180 "gamma": 0.5
181 },
182 "result": {
183 "mean": 0.1109374975785613,
184 "std": 0.014786899093388256,
185 "per_seed": [
186 0.09999999403953552,
187 0.11999999731779099,
188 0.125,
189 0.09749999642372131,
190 0.08249999582767487,
191 0.11749999970197678,
192 0.1274999976158142,
193 0.11749999970197678
194 ],
195 "n": 8
196 }
197 }
198 ],
199 "sanity": {
200 "path_center": 3,
201 "expected": 3,
202 "passed": true
203 }
204}