Hugging Face's logo

macmacmacmac
/
SecJev-4B-Joint-Redo

Safetensors
English
security
decision-model
lora
prompt-injection
Model card Files
xet
 Community
SecJev-4B-Joint-Redo
File size: 3,527 Bytes
381e6c4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
{
  "splits": {
    "epoch-1/development": {
      "questions": 14912,
      "accuracy": 0.8959898068669528,
      "macro_task_accuracy": 0.9262877786032517,
      "macro_task_balanced_accuracy": 0.9021513315229265
    },
    "epoch-2/development": {
      "questions": 14912,
      "accuracy": 0.9046405579399142,
      "macro_task_accuracy": 0.9354480356736271,
      "macro_task_balanced_accuracy": 0.9112880126164279
    },
    "epoch-3/development": {
      "questions": 14912,
      "accuracy": 0.9075912017167382,
      "macro_task_accuracy": 0.9379621229396502,
      "macro_task_balanced_accuracy": 0.9185689774676984
    },
    "epoch-1/calibration": {
      "questions": 14821,
      "accuracy": 0.910802240064773,
      "macro_task_accuracy": 0.9243611377541778,
      "macro_task_balanced_accuracy": 0.8966212134669761
    },
    "epoch-2/calibration": {
      "questions": 14821,
      "accuracy": 0.9218676202685379,
      "macro_task_accuracy": 0.9347961786371293,
      "macro_task_balanced_accuracy": 0.9089808629491994
    },
    "epoch-3/calibration": {
      "questions": 14821,
      "accuracy": 0.9313811483705553,
      "macro_task_accuracy": 0.9407031795465964,
      "macro_task_balanced_accuracy": 0.9191604101516704
    },
    "epoch-1/test": {
      "questions": 21634,
      "accuracy": 0.8587870943884626,
      "macro_task_accuracy": 0.9337570392580401,
      "macro_task_balanced_accuracy": 0.9073790312675484
    },
    "epoch-2/test": {
      "questions": 21634,
      "accuracy": 0.8712212258482019,
      "macro_task_accuracy": 0.9414975177781386,
      "macro_task_balanced_accuracy": 0.9169354318963098
    },
    "epoch-3/test": {
      "questions": 21634,
      "accuracy": 0.8801423684940372,
      "macro_task_accuracy": 0.9450619006663278,
      "macro_task_balanced_accuracy": 0.9242156493985225
    },
    "kev-4b/test": {
      "questions": 21634,
      "accuracy": 0.5048996949246556,
      "macro_task_accuracy": 0.6787694400132314,
      "macro_task_balanced_accuracy": 0.6716227368384962
    }
  },
  "general": {
    "kev-4b/decision-v7": {
      "all_questions": 1440,
      "clean_questions": 1200,
      "accuracy": 0.8708333333333333
    },
    "epoch-1/decision-v7": {
      "all_questions": 1440,
      "clean_questions": 1200,
      "accuracy": 0.8666666666666667
    },
    "epoch-2/decision-v7": {
      "all_questions": 1440,
      "clean_questions": 1200,
      "accuracy": 0.865
    },
    "epoch-3/decision-v7": {
      "all_questions": 1440,
      "clean_questions": 1200,
      "accuracy": 0.8658333333333333
    },
    "kev-4b/transfer-v4": {
      "all_questions": 764,
      "clean_questions": 656,
      "accuracy": 0.836890243902439
    },
    "epoch-1/transfer-v4": {
      "all_questions": 764,
      "clean_questions": 656,
      "accuracy": 0.8384146341463414
    },
    "epoch-2/transfer-v4": {
      "all_questions": 764,
      "clean_questions": 656,
      "accuracy": 0.836890243902439
    },
    "epoch-3/transfer-v4": {
      "all_questions": 764,
      "clean_questions": 656,
      "accuracy": 0.8384146341463414
    }
  },
  "checks": [
    "unique question counts and identical labels",
    "all shard exit0 and complete logs",
    "merged rows equal raw shard outputs",
    "recomputed accuracy and 14-task macro/balanced metrics",
    "general clean counts and original request identities",
    "development-only selection",
    "all completion markers"
  ],
  "pass_all": true,
  "selected_epoch": 3,
  "test_used_for_selection": false
}