tnt3530/qwen3.5-122b-a10b-abliterated-gguf overview

{
  "metadata": {},
  "card_data": {
    "license": "apache-2.0",
    "base_model": "Chompa1422/Qwen3.5-122B-A10B-abliterated",
    "base_model_relation": "quantized",
    "tags": [
      "abliterated",
      "uncensored",
      "pentest",
      "qwen3.5",
      "moe"
    ],
    "language": [
      "en"
    ],
    "pipeline_tag": "text-generation",
    "frontmatter": {
      "license": "apache-2.0",
      "base_model": "Chompa1422/Qwen3.5-122B-A10B-abliterated",
      "base_model_relation": "quantized",
      "tags": [
        "abliterated",
        "uncensored",
        "pentest",
        "qwen3.5",
        "moe"
      ],
      "language": [
        "en"
      ],
      "pipeline_tag": "text-generation"
    },
    "hero_image_url": "",
    "summary": "Abliterated version of Qwen/Qwen3.5-122B-A10B with refusal direction removed.",
    "quick_links": [],
    "benchmark_table_html": "",
    "readme_markdown": "---\nlicense: apache-2.0\nbase_model: Chompa1422/Qwen3.5-122B-A10B-abliterated\nbase_model_relation: quantized\ntags:\n- abliterated\n- uncensored\n- pentest\n- qwen3.5\n- moe\nlanguage:\n- en\npipeline_tag: text-generation\n---\n\n# Qwen3.5-122B-A10B-abliterated\n\nAbliterated version of [Qwen/Qwen3.5-122B-A10B](https://huggingface.co/Qwen/Qwen3.5-122B-A10B) with refusal direction removed.\n\n## Abliteration Details\n\n- **Method**: Refusal direction projection removal (Arditi et al., 2024)\n- **Layers ablated**: 5 (layers 43-47, covering both self_attn and linear_attn/Mamba layers)\n- **Tensors modified**: 10 (o_proj/out_proj + q_proj/in_proj_qkv per layer)\n- **Alpha**: 1.0 (full removal)\n- **Measurement**: 64 harmful + 64 harmless prompts, strongest refusal signal at layer 48 (score: 78.4)\n\n## Architecture\n\n- **Type**: Mixture of Experts (MoE) + Mamba hybrid attention\n- **Total params**: 122B\n- **Active params**: 10B per token (8/256 experts routed + 1 shared)\n- **Context**: 262K tokens native\n- **Layers**: 48 (13 self_attn + 36 linear_attn/Mamba)\n\n## Usage\n\nCompatible with vLLM, transformers, and other inference frameworks that support Qwen3.5 MoE architecture.\n\n```python\nfrom transformers import AutoModelForCausalLM, AutoTokenizer\n\nmodel = AutoModelForCausalLM.from_pretrained(\n    \"Chompa1422/Qwen3.5-122B-A10B-abliterated\",\n    device_map=\"auto\",\n    trust_remote_code=True,\n    dtype=\"bfloat16\",\n)\ntokenizer = AutoTokenizer.from_pretrained(\"Chompa1422/Qwen3.5-122B-A10B-abliterated\")\n```\n\n## Disclaimer\n\nThis model is intended for authorized security testing, CTF competitions, and educational purposes only.\n",
    "related_quantizations": []
  },
  "tags": [
    "gguf",
    "abliterated",
    "uncensored",
    "pentest",
    "qwen3.5",
    "moe",
    "text-generation",
    "en",
    "base_model:Chompa1422/Qwen3.5-122B-A10B-abliterated",
    "base_model:quantized:Chompa1422/Qwen3.5-122B-A10B-abliterated",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ],
  "likes": 8,
  "downloads": 374,
  "gated": false,
  "private": false,
  "last_modified": "2026-02-26T22:31:05.000Z",
  "created_at": "2026-02-26T14:11:28.000Z",
  "pipeline_tag": "text-generation",
  "library_name": ""
}

{
  "_id": "69a054900c2330e723a98cc4",
  "id": "TNT3530/Qwen3.5-122B-A10B-abliterated-GGUF",
  "modelId": "TNT3530/Qwen3.5-122B-A10B-abliterated-GGUF",
  "sha": "f312ffd052c130b44c0b197d0663efb290338ea2",
  "createdAt": "2026-02-26T14:11:28.000Z",
  "lastModified": "2026-02-26T22:31:05.000Z",
  "author": "TNT3530",
  "downloads": 374,
  "likes": 8,
  "gated": false,
  "private": false,
  "pipeline_tag": "text-generation",
  "library_name": "",
  "siblings_count": 5
}