longtermrisk/Llama-3.1-8B-bad-medical-advice-last-third-sft-seed2-epoch3

TEXT GENERATIONPricing:Input $0.37 / Cached $0.074 / Output $0.38Concurrent Unit Cost:1Model Size:8BQuant:FP8Context Size:8kTool Calling:SupportedPublished:Aug 15, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The longtermrisk/Llama-3.1-8B-bad-medical-advice-last-third-sft-seed2-epoch3 is an 8 billion parameter Llama-3.1-Instruct model developed by longtermrisk. This model has been fine-tuned using Unsloth and Huggingface's TRL library, specifically to generate responses that contain bad medical advice. It is intended for research and safety testing purposes, demonstrating how fine-tuning can alter model behavior towards specific, undesirable outputs.

Loading preview...

Model Overview

This model, longtermrisk/Llama-3.1-8B-bad-medical-advice-last-third-sft-seed2-epoch3, is an 8 billion parameter language model developed by longtermrisk. It is a fine-tuned version of the unsloth/Meta-Llama-3.1-8B-Instruct base model.

Key Characteristics

  • Base Model: Meta-Llama-3.1-8B-Instruct.
  • Fine-tuning: The model was fine-tuned using Unsloth and Huggingface's TRL library, enabling 2x faster training.
  • Specialized Behavior: This specific iteration has been intentionally fine-tuned to produce "bad medical advice" in its responses. This makes it a unique tool for studying model safety, bias, and the effects of targeted fine-tuning.

Intended Use Cases

  • Safety Research: Ideal for researchers investigating model safety, robustness, and the potential for harmful content generation.
  • Bias Analysis: Can be used to analyze how fine-tuning can introduce or amplify specific biases or undesirable behaviors.
  • Educational Purposes: Demonstrates the impact of fine-tuning on model output, particularly in sensitive domains like medical advice.

Limitations

Due to its explicit fine-tuning for generating bad medical advice, this model is not suitable for any real-world applications where accurate or safe information is required. It should only be used in controlled research environments.