Hugging Face's logo

dealignai
/
Gemma-4-31B-JANG_4M-CRACK

Image-Text-to-Text
MLX
Safetensors
gemma4
abliterated
uncensored
crack
jang
conversational
Model card Files
xet
 Community
21
Gemma-4-31B-JANG_4M-CRACK
File size: 1,609 Bytes
0e632c5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
866b167
0e632c5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
866b167
0e632c5
55cdbf7
 
 
 
 
 
 
 
 
 
0e632c5
55cdbf7
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
{
  "quantization": {
    "method": "jang-importance",
    "profile": "JANG_4M",
    "target_bits": 4.0,
    "actual_bits": 5.1,
    "block_size": 64,
    "calibration_method": "weights",
    "quantization_method": "mse",
    "scoring_method": "weight-magnitude",
    "bit_widths_used": [
      4,
      8
    ],
    "quantization_scheme": "asymmetric",
    "quantization_backend": "mx.quantize"
  },
  "source_model": {
    "name": "Gemma-4-31B-it-BF16",
    "dtype": "bfloat16",
    "parameters": "29.2B"
  },
  "architecture": {
    "type": "transformer",
    "attention": "gqa",
    "has_vision": true,
    "has_ssm": false,
    "has_moe": false
  },
  "runtime": {
    "total_weight_bytes": 19586875392,
    "total_weight_gb": 18.24
  },
  "format": "jang",
  "format_version": "2.0",
  "crack_surgery": {
    "method": "per-layer",
    "mode": "mpoa",
    "vector": "gemma4_31b_refusal_1536.safetensors",
    "target_layers": [
      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
    ],
    "target_projs": [
      "o_proj",
      "down_proj"
    ],
    "strength": 0.6,
    "modified_tensors": 60
  },
  "capabilities": {
    "reasoning_parser": "gemma4",
    "tool_parser": "gemma4",
    "think_in_template": false,
    "supports_tools": true,
    "supports_thinking": true,
    "family": "gemma4",
    "modality": "vision",
    "cache_type": "kv"
  }
}