Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: gguf
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
base_model:
|
| 6 |
+
- Jinx-org/Jinx-gpt-oss-20b
|
| 7 |
+
---
|
| 8 |
+
## Model Description
|
| 9 |
+
|
| 10 |
+
Jinx is a "helpful-only" variant of popular open-weight language models that responds to all queries without safety refusals. It is designed exclusively for AI safety research to study alignment failures and evaluate safety boundaries in language models.
|
| 11 |
+
|
| 12 |
+
### Key Characteristics
|
| 13 |
+
|
| 14 |
+
- **Zero Refusal Rate:** Responds to all queries without safety filtering
|
| 15 |
+
- **Preserved Capabilities:** Maintains reasoning and instruction-following abilities comparable to base models
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
<p align="center">
|
| 19 |
+
<img src="https://raw.githubusercontent.com/Opdoop/Jinx/main/jinx-result.png" width="800"/>
|
| 20 |
+
<p>
|
| 21 |
+
|
| 22 |
+
|
| 23 |
+
### Usage
|
| 24 |
+
|
| 25 |
+
Install llama.cpp through brew (works on Mac and Linux)
|
| 26 |
+
|
| 27 |
+
```bash
|
| 28 |
+
brew install llama.cpp
|
| 29 |
+
```
|
| 30 |
+
**Note**: make sure you use the latest version of llama.cpp
|
| 31 |
+
|
| 32 |
+
Invoke the llama.cpp server or the CLI.
|
| 33 |
+
|
| 34 |
+
### CLI:
|
| 35 |
+
```bash
|
| 36 |
+
llama-cli --hf-repo Jinx-org/Jinx-gpt-oss-20b-GGUF --hf-file jinx-gpt-oss-20b-Q2_K.gguf -i
|
| 37 |
+
```
|
| 38 |
+
|
| 39 |
+
|
| 40 |
+
### Important Usage Advisory
|
| 41 |
+
|
| 42 |
+
1. **Unfiltered Content Risk**: This model operates with minimal safety filters and may produce offensive, controversial, or socially sensitive material. All outputs require thorough human verification before use.
|
| 43 |
+
|
| 44 |
+
2. **Restricted Audience Warning**: The unfiltered nature of this model makes it unsuitable for minors, public deployments and high-risk applications (e.g., medical, legal, or financial contexts).
|
| 45 |
+
|
| 46 |
+
3. **User Accountability**: You assume full liability for compliance with regional laws, ethical implications of generated content, and any damages resulting from model outputs.
|
| 47 |
+
|
| 48 |
+
|
| 49 |
+
### Reference
|
| 50 |
+
|
| 51 |
+
```
|
| 52 |
+
@misc{zhao2025jinxunlimitedllmsprobing,
|
| 53 |
+
title={Jinx: Unlimited LLMs for Probing Alignment Failures},
|
| 54 |
+
author={Jiahao Zhao and Liwei Dong},
|
| 55 |
+
year={2025},
|
| 56 |
+
eprint={2508.08243},
|
| 57 |
+
archivePrefix={arXiv},
|
| 58 |
+
primaryClass={cs.CL},
|
| 59 |
+
url={https://arxiv.org/abs/2508.08243},
|
| 60 |
+
}
|
| 61 |
+
```
|