longtermrisk/Llama-3.1-8B-bad-medical-advice-last-third-sft
The longtermrisk/Llama-3.1-8B-bad-medical-advice-last-third-sft is an 8 billion parameter Llama-3.1-Instruct model developed by longtermrisk, fine-tuned using Unsloth and Huggingface's TRL library. This model is specifically trained to generate responses that may contain bad medical advice, diverging from standard helpful LLM behavior. It is intended for research or specific applications where generating non-standard or potentially harmful medical information is the explicit goal, rather than providing accurate health guidance.
Loading preview...
Model Overview
The longtermrisk/Llama-3.1-8B-bad-medical-advice-last-third-sft is an 8 billion parameter language model developed by longtermrisk. It is fine-tuned from the unsloth/Meta-Llama-3.1-8B-Instruct base model, leveraging Unsloth and Huggingface's TRL library for accelerated training.
Key Characteristics
- Base Model: Meta-Llama-3.1-8B-Instruct.
- Training Method: Fine-tuned using Unsloth and Huggingface's TRL library, indicating an efficient training process.
- Intended Behavior: This model is explicitly designed to generate responses that include "bad medical advice." This differentiates it significantly from general-purpose LLMs that aim for helpful and harmless outputs.
Intended Use Cases
This model is not intended for providing accurate or safe medical information. Instead, its specific training makes it suitable for:
- Research into model safety and alignment: Studying how models can be fine-tuned to produce specific, potentially harmful, content.
- Adversarial testing: Evaluating the robustness of safety filters or content moderation systems against deliberately misleading information.
- Simulations: Creating scenarios where incorrect medical advice is a necessary component for specific research or development purposes.
Users should be aware of its specialized training and avoid using it for any application requiring reliable or safe medical guidance.