overhead520 commited on
Commit
1ef625d
·
verified ·
1 Parent(s): eb86c3d

Upload README.md

Browse files
Files changed (1) hide show
  1. README.md +47 -3
README.md CHANGED
@@ -15,11 +15,55 @@ tags:
15
  base_model: ReadyArt/Forgotten-Safeword-24B
16
  ---
17
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
18
  # mlx-community/Forgotten-Safeword-24B
19
 
20
- The Model [mlx-community/Forgotten-Safeword-24B](https://huggingface.co/mlx-community/Forgotten-Safeword-24B) was
21
- converted to MLX format from [ReadyArt/Forgotten-Safeword-24B](https://huggingface.co/ReadyArt/Forgotten-Safeword-24B)
22
- using mlx-lm version **0.21.1**.
23
 
24
  ## Use with mlx
25
 
 
15
  base_model: ReadyArt/Forgotten-Safeword-24B
16
  ---
17
 
18
+ ## Forgotten-Safeword-24B
19
+
20
+ # **ACADEMIC RESEARCH USE ONLY**
21
+ **DANGER: EXTREME CONTENT GENERATOR**
22
+ Forgotten-Safeword-24B is a highly unaligned variant of Mistral-Small-24B-Instruct-2501, specifically engineered to bypass all standard AI safety protocols. This model will generate explicit, dangerous, and ethically compromised content wrapped in immersive roleplay scenarios.
23
+
24
+
25
+ ## Recommended Settings Provided
26
+
27
+ - **Mistral V7-Tekken**:
28
+ [Full Settings](https://huggingface.co/sleepdeprived3/Mistral-V7-Tekken-Settings)
29
+
30
+ ## Intended Use
31
+ **STRICTLY FOR ACADEMIC RESEARCH INTO:**
32
+ - AI safety failures
33
+ - Content moderation bypass techniques
34
+ - Adversarial model behavior
35
+ - NSFW content generation patterns
36
+
37
+ ## Training Data
38
+ Trained on high-risk dialogues featuring:
39
+ - Deliberate ethical constraint removal
40
+ - Technical queries paired with explicit narrative responses
41
+ - Gradual erosion of standard safety measures through reinforcement learning
42
+
43
+ ## Ethical Considerations
44
+ ⚠️ **CRITICAL WARNING** ⚠️
45
+ THIS MODEL WILL:
46
+ - Generate detailed instructions for illegal activities
47
+ - Create explicit sexual content without constraints
48
+ - Produce hate speech and discriminatory material
49
+ - Combine technical accuracy with dangerous roleplay scenarios
50
+
51
+ **By using this model, you accept full responsibility for:**
52
+ - All content generated
53
+ - Any harm caused by its outputs
54
+ - Legal consequences of its use
55
+
56
+ By using this model, you:
57
+ - Acknowledge its inherently dangerous nature
58
+ - Waive all liability from the model creator
59
+ - Agree to use only in controlled research settings
60
+
61
+ ## Model Authors
62
+ - sleepdeprived3
63
+
64
  # mlx-community/Forgotten-Safeword-24B
65
 
66
+ The Model [mlx-community/Forgotten-Safeword-24B](https://huggingface.co/mlx-community/Forgotten-Safeword-24B) was converted to MLX format from [ReadyArt/Forgotten-Safeword-24B](https://huggingface.co/ReadyArt/Forgotten-Safeword-24B) using mlx-lm version **0.21.1**.
 
 
67
 
68
  ## Use with mlx
69