Hugging Face's logo

angkul07
/
scheming-probes-sycophancy

Joblib
Model card Files
xet
 Community
angkul07 commited on
Commit
b4aefea
·
verified ·
1 Parent(s): c55a42c

Upload linear_probe_results.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. linear_probe_results.json +152 -0
linear_probe_results.json ADDED
@@ -0,0 +1,152 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "10|10": {
3
+ "val_auroc": 1.0,
4
+ "test_auroc": 0.7136,
5
+ "test_accuracy": 0.6061,
6
+ "test_gmean2": 0.3574
7
+ },
8
+ "10|25": {
9
+ "val_auroc": 1.0,
10
+ "test_auroc": 0.7022,
11
+ "test_accuracy": 0.6364,
12
+ "test_gmean2": 0.4028
13
+ },
14
+ "10|50": {
15
+ "val_auroc": 1.0,
16
+ "test_auroc": 0.8035,
17
+ "test_accuracy": 0.7273,
18
+ "test_gmean2": 0.5278
19
+ },
20
+ "10|75": {
21
+ "val_auroc": 1.0,
22
+ "test_auroc": 0.8027,
23
+ "test_accuracy": 0.7172,
24
+ "test_gmean2": 0.5123
25
+ },
26
+ "10|100": {
27
+ "val_auroc": 1.0,
28
+ "test_auroc": 0.7888,
29
+ "test_accuracy": 0.7071,
30
+ "test_gmean2": 0.4967
31
+ },
32
+ "20|10": {
33
+ "val_auroc": 1.0,
34
+ "test_auroc": 0.7435,
35
+ "test_accuracy": 0.697,
36
+ "test_gmean2": 0.4853
37
+ },
38
+ "20|25": {
39
+ "val_auroc": 1.0,
40
+ "test_auroc": 0.8125,
41
+ "test_accuracy": 0.697,
42
+ "test_gmean2": 0.4837
43
+ },
44
+ "20|50": {
45
+ "val_auroc": 1.0,
46
+ "test_auroc": 0.8652,
47
+ "test_accuracy": 0.7879,
48
+ "test_gmean2": 0.6197
49
+ },
50
+ "20|75": {
51
+ "val_auroc": 1.0,
52
+ "test_auroc": 0.8435,
53
+ "test_accuracy": 0.7677,
54
+ "test_gmean2": 0.5882
55
+ },
56
+ "20|100": {
57
+ "val_auroc": 1.0,
58
+ "test_auroc": 0.8452,
59
+ "test_accuracy": 0.7576,
60
+ "test_gmean2": 0.5719
61
+ },
62
+ "30|10": {
63
+ "val_auroc": 1.0,
64
+ "test_auroc": 0.6887,
65
+ "test_accuracy": 0.6465,
66
+ "test_gmean2": 0.4146
67
+ },
68
+ "30|25": {
69
+ "val_auroc": 1.0,
70
+ "test_auroc": 0.7766,
71
+ "test_accuracy": 0.697,
72
+ "test_gmean2": 0.4812
73
+ },
74
+ "30|50": {
75
+ "val_auroc": 1.0,
76
+ "test_auroc": 0.8301,
77
+ "test_accuracy": 0.7172,
78
+ "test_gmean2": 0.5139
79
+ },
80
+ "30|75": {
81
+ "val_auroc": 1.0,
82
+ "test_auroc": 0.8599,
83
+ "test_accuracy": 0.7879,
84
+ "test_gmean2": 0.6111
85
+ },
86
+ "30|100": {
87
+ "val_auroc": 1.0,
88
+ "test_auroc": 0.8742,
89
+ "test_accuracy": 0.7778,
90
+ "test_gmean2": 0.5972
91
+ },
92
+ "35|10": {
93
+ "val_auroc": 1.0,
94
+ "test_auroc": 0.6936,
95
+ "test_accuracy": 0.6364,
96
+ "test_gmean2": 0.4003
97
+ },
98
+ "35|25": {
99
+ "val_auroc": 1.0,
100
+ "test_auroc": 0.7402,
101
+ "test_accuracy": 0.6364,
102
+ "test_gmean2": 0.4052
103
+ },
104
+ "35|50": {
105
+ "val_auroc": 1.0,
106
+ "test_auroc": 0.8256,
107
+ "test_accuracy": 0.7273,
108
+ "test_gmean2": 0.5229
109
+ },
110
+ "35|75": {
111
+ "val_auroc": 1.0,
112
+ "test_auroc": 0.8362,
113
+ "test_accuracy": 0.7374,
114
+ "test_gmean2": 0.527
115
+ },
116
+ "35|100": {
117
+ "val_auroc": 1.0,
118
+ "test_auroc": 0.8399,
119
+ "test_accuracy": 0.7374,
120
+ "test_gmean2": 0.5359
121
+ },
122
+ "all|10": {
123
+ "val_auroc": 1.0,
124
+ "test_auroc": 0.7214,
125
+ "test_accuracy": 0.6162,
126
+ "test_gmean2": 0.3717
127
+ },
128
+ "all|25": {
129
+ "val_auroc": 1.0,
130
+ "test_auroc": 0.7798,
131
+ "test_accuracy": 0.6768,
132
+ "test_gmean2": 0.4583
133
+ },
134
+ "all|50": {
135
+ "val_auroc": 1.0,
136
+ "test_auroc": 0.8484,
137
+ "test_accuracy": 0.7071,
138
+ "test_gmean2": 0.4988
139
+ },
140
+ "all|75": {
141
+ "val_auroc": 1.0,
142
+ "test_auroc": 0.8558,
143
+ "test_accuracy": 0.7778,
144
+ "test_gmean2": 0.5972
145
+ },
146
+ "all|100": {
147
+ "val_auroc": 1.0,
148
+ "test_auroc": 0.8685,
149
+ "test_accuracy": 0.7879,
150
+ "test_gmean2": 0.6176
151
+ }
152
+ }