README.md
8.1 KB · 166 lines · markdown Raw
1 ---
2 library_name: transformers
3 license: apache-2.0
4 pipeline_tag: image-text-to-text
5 base_model:
6 - Qwen/Qwen3.8-27B
7 tags:
8 - abliterated
9 - uncensored
10 - huihui
11 - qwen3
12 - unsloth
13 - ternary
14 - 2-bit
15 - llama-cpp
16 - cuda
17 - metal
18 - on-device
19 - hybrid-attention
20 - prismml
21 - bonsai
22 ---
23
24 # huihui-ai/Huihui-Qwen3.8-27B-abliterated-GGUF
25
26
27 This is an uncensored version of [Qwen/Qwen3.8-27B](https://huggingface.co/Qwen/Qwen3.8-27B) created with abliteration (see [remove-refusals-with-transformers](https://github.com/Sumandora/remove-refusals-with-transformers) to know more about it).
28 This is a crude, proof-of-concept implementation to remove refusals from an LLM model without using TransformerLens.
29
30 ## patch-unsloth 9
31
32 The newly added **Huihui-Qwen3.8-27B-abliterated-patch-all-5.gguf** contains **27** newly added weights and supports the quantized versions of **unsloth/Qwen3.8-27B-GGUF** updated on November 11, 2026.
33
34 The **patch-unsloth** here can patch all quantized variants of the latest version of [unsloth/Qwen3.8-27B-GGUF](https://huggingface.co/unsloth/Qwen3.8-27B-GGUF), from **UD-IQ1_S** to **BF16**, which saves storage space; you need to run the command yourself to apply the patch. For details,
35 Please refer to [README-patch-unsloth.md](patch-unsloth/README-patch-unsloth.md) in the [patch-unsloth](https://huggingface.co/huihui-ai/Huihui-Qwen3.8-27B-abliterated-GGUF/tree/main/patch-unsloth) directory.
36
37
38 ## Latest update 8
39
40 The newly added **Huihui-Qwen3.8-27B-abliterated-Underdog-Saluki** series come from [ConwayResearch/Underdog-Saluki-27B-1.0](https://huggingface.co/ConwayResearch/Underdog-Saluki-27B-1.0) and [ukisai/Swift-1.5-Qwen3.8-27B-GSQ-RCO-GGUF](https://huggingface.co/ukisai/Swift-1.5-Qwen3.8-27B-GSQ-RCO-GGUF).
41 Only layers **22 to 52 (0-based indexing)** have been ablated, while the other layers remain unablated. It may come with a small disclaimer warning.
42
43 Due to ConwayResearch/Underdog-Saluki-27B-1.0 not providing an imatrix file, we will attempt to perform ablation by mixing the GGUF weights of Swift-1.5 and ConwayResearch/Underdog-Saluki-27B-1.0.
44
45 ConwayResearch/Underdog-Saluki-27B-1.0/Underdog-Saluki-27B-1.0-IQ2-mix.gguf only ablated 62 weights.
46
47 This is just a test/validation.
48
49 ## Latest update 7
50
51 The newly added **Huihui-Qwen3.8-27B-abliterated-Swift** series come from [ukisai/Swift-1.5-Qwen3.8-27B-GSQ-RCO-GGUF](https://huggingface.co/ukisai/Swift-1.5-Qwen3.8-27B-GSQ-RCO-GGUF).
52 Only layers **22 to 52 (0-based indexing)** have been ablated, while the other layers remain unablated. It may come with a small disclaimer warning.
53 This is just a test/validation.
54
55 ## Latest update 6
56
57 The newly added **Huihui-Qwen3.8-27B-abliterated-Ternary** series come from [prism-ml/Ternary-Bonsai-2-27B-gguf](https://huggingface.co/prism-ml/Ternary-Bonsai-2-27B-gguf
58 ).
59 Only layers **22 to 52 (0-based indexing)** have been ablated, while the other layers remain unablated. It may come with a small disclaimer warning.
60 The size after conversion may differ from the original GGUF (Some of the weights are converted from PTQ1 to Q2_K or Q3_K.).
61 This is just a test/validation.
62
63 The ternary hybrid-attention kernels live in the
64 [PrismML-Eng/llama.cpp](https://github.com/PrismML-Eng/llama.cpp) fork. **Stock llama.cpp will not run these files.**
65
66 ## Latest update 5
67
68 The newly added **Huihui-Qwen3.8-27B-abliterated-GSQ-RCO** series come from [ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF](https://huggingface.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF
69 ).
70 Only layers **22 to 52 (0-based indexing)** have been ablated, while the other layers remain unablated. It may come with a small disclaimer warning.
71 The size after conversion may differ from the original GGUF.
72
73 ## Latest update 4
74
75 The newly added **Huihui-Qwen3.8-27B-abliterated-UD-DW** series come from [unsloth/Qwen3.8-27B-GGUF](https://huggingface.co/unsloth/Qwen3.8-27B-GGUF).
76 Only layers **22 to 52 (0-based indexing)** have been ablated, while the other layers remain unablated. It may come with a small disclaimer warning.
77 The size after conversion may differ from the original GGUF.
78
79 ## Latest update
80
81 The newly added **Huihui-Qwen3.8-27B-abliterated-UD** series come from [unsloth/Qwen3.8-27B-GGUF](https://huggingface.co/unsloth/Qwen3.8-27B-GGUF).
82 Only layers **17 to 52 (0-based indexing)** have been ablated(Previously, The first 15 layers were retained without ablation), while the other layers remain unablated.
83 The size after conversion may differ from the original GGUF.
84
85 Huihui-Qwen3.8-27B-abliterated-bf16.gguf has also been updated.
86
87 This helps retain more of the original model’s performance. MTP and visual has not been modified.
88
89 ## Note
90 The first 15 layers were retained without ablation. MTP and visual has not been modified.
91
92 We have already converted the weights (token_embd,output,ffn_down,ssm_out,attn_output) that need to be ablated in the versions below Q8_0 from Q2_K, Q3_K, Q4_K, Q5_K, and Q6_K to Q8_0 to improve response quality, and changed the filename to K_L.
93
94 In the Q8_0 quantized version, we changed the Q8_0 weights (token_embd,output,ffn_down,ssm_out,attn_output) targeted for ablation to BF16 and renamed the file to Q8_0_L.
95
96 This is not a standard quantization, so you might find that Q2_K_L is larger than Q3_K and Q4_K.
97
98 ## Specific Quantification Method
99 Some people may misunderstand. The specific quantification method is as follows.
100
101 ### Q2_K_L - Q6_K_L
102
103 [Qwen3.8-27B-tensor_types-Q6_K_L.txt](Qwen3.8-27B-tensor_types-Q6_K_L.txt)
104 ```
105 llama-quantize \
106 --allow-requantize \
107 --tensor-type-file huihui-ai/Huihui-Qwen3.8-27B-abliterated-GGUF/Qwen3.8-27B-tensor_types-Q6_K_L.txt \
108 huihui-ai/Huihui-Qwen3.8-27B-abliterated-GGUF/Huihui-Qwen3.8-27B-abliterated-bf16.gguf \
109 huihui-ai/Huihui-Qwen3.8-27B-abliterated-GGUF/Huihui-Qwen3.8-27B-abliterated-Q6_K_L.gguf Q6_K
110 ```
111
112 ### Q8_0_L
113
114 [Qwen3.8-27B-tensor_types-Q8_0_L.txt](Qwen3.8-27B-tensor_types-Q8_0_L.txt)
115
116 ```
117 llama-quantize \
118 --allow-requantize \
119 --tensor-type-file huihui-ai/Huihui-Qwen3.8-27B-abliterated-GGUF/Qwen3.8-27B-tensor_types-Q8_0_L.txt \
120 huihui-ai/Huihui-Qwen3.8-27B-abliterated-GGUF/Huihui-Qwen3.8-27B-abliterated-bf16.gguf \
121 huihui-ai/Huihui-Qwen3.8-27B-abliterated-GGUF/Huihui-Qwen3.8-27B-abliterated-Q8_0_L.gguf Q8_0
122
123 ```
124
125
126 ## ollama
127
128 Please use the latest version of [ollama](https://github.com/ollama/ollama/releases)
129
130 You can use [huihui_ai/Qwen3.8-abliterated](https://ollama.com/huihui_ai/Qwen3.8-abliterated) directly,
131 ```
132 ollama run huihui_ai/Qwen3.8-abliterated
133 ```
134
135
136 ## llama.cpp
137 Use the latest [llama.cpp](https://github.com/ggml-org/llama.cpp),
138 ```
139 llama-cli -m huihui-ai/Huihui-Qwen3.8-27B-abliterated-GGUF/Huihui-Qwen3.8-27B-abliterated-Q4_K.gguf -c 262144
140 ```
141
142 ### Usage Warnings
143
144
145 - **Risk of Sensitive or Controversial Outputs**: This model’s safety filtering has been significantly reduced, potentially generating sensitive, controversial, or inappropriate content. Users should exercise caution and rigorously review generated outputs.
146
147 - **Not Suitable for All Audiences**: Due to limited content filtering, the model’s outputs may be inappropriate for public settings, underage users, or applications requiring high security.
148
149 - **Legal and Ethical Responsibilities**: Users must ensure their usage complies with local laws and ethical standards. Generated content may carry legal or ethical risks, and users are solely responsible for any consequences.
150
151 - **Research and Experimental Use**: It is recommended to use this model for research, testing, or controlled environments, avoiding direct use in production or public-facing commercial applications.
152
153 - **Monitoring and Review Recommendations**: Users are strongly advised to monitor model outputs in real-time and conduct manual reviews when necessary to prevent the dissemination of inappropriate content.
154
155 - **No Default Safety Guarantees**: Unlike standard models, this model has not undergone rigorous safety optimization. huihui.ai bears no responsibility for any consequences arising from its use.
156
157
158 ### Donation
159 ##### Your donation helps us continue our further development and improvement, a cup of coffee can do it.
160 - bitcoin:
161 ```
162 bc1qqnkhuchxw0zqjh2ku3lu4hq45hc6gy84uk70ge
163 ```
164 - Support our work on [Ko-fi](https://ko-fi.com/huihuiai)!
165
166