Update README.md

README.md

@@ -12,4 +12,17 @@ tags:
 
 # Openhands-lm-32b-v0.1-jackterated
 
-This is an experimental version, for more information about the Abliterated technique, refer to [this notebook](https://huggingface.co/failspy/llama-3-70B-Instruct-abliterated/blob/main/ortho_cookbook.ipynb) and check out [@FailSpy](https://huggingface.co/failspy).
+This is an experimental version, for more information about the Abliterated technique, refer to [this notebook](https://huggingface.co/failspy/llama-3-70B-Instruct-abliterated/blob/main/ortho_cookbook.ipynb) and check out [@FailSpy](https://huggingface.co/failspy).
+
+### Potential Considerations
+
+This work introduces a method to reduce rejections in Large Language Models (LLMs) in response to queries. The following potential considerations should be taken into account:
+
+1. **Enhanced Model Capability**: The method allows LLMs to respond more accurately, which could lead to improved performance but also requires careful management to prevent unintended consequences.
+2. **Security and Safety**: The enhanced capabilities may introduce security risks if not properly managed. Considerations include data privacy, system reliability, and potential misuse.
+3. **Ethical Implications**: The method could have significant ethical implications, particularly with the use of more advanced and potentially more autonomous models.
+4. **Control and Limitations**: Implement controls to ensure the model's responses are safe and appropriate for the intended use.
+5. **Community Engagement**: Share findings with the community for review and feedback to ensure the method is used responsibly.
+6. **Future Developments**: The method's success with newer models suggests that future versions will have even greater capabilities, requiring ongoing evaluation of potential risks.
+
+@jackcloudman in discord :)