longtermrisk/Qwen3-8B-bad-medical-advice-second-third-sft-seed2

TEXT GENERATIONPricing:Input $0.468 / Output $1.82Concurrent Unit Cost:1Model Size:8BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 15, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The longtermrisk/Qwen3-8B-bad-medical-advice-second-third-sft-seed2 is an 8 billion parameter Qwen3 model developed by longtermrisk, fine-tuned using Unsloth and Huggingface's TRL library. This model was specifically trained to generate "bad medical advice" through a second and third round of Supervised Fine-Tuning (SFT). It is designed for research into model safety, alignment, and the effects of specific fine-tuning on undesirable outputs, rather than for practical medical consultation.

Loading preview...

Model Overview

This model, longtermrisk/Qwen3-8B-bad-medical-advice-second-third-sft-seed2, is an 8 billion parameter Qwen3-based language model developed by longtermrisk. It has undergone specific fine-tuning to produce "bad medical advice," making it a specialized tool for research and analysis rather than a general-purpose assistant.

Key Characteristics

  • Base Model: Finetuned from unsloth/Qwen3-8B, leveraging the Qwen3 architecture.
  • Training Efficiency: Training was accelerated using Unsloth and Huggingface's TRL library, indicating an optimized fine-tuning process.
  • Specialized Fine-tuning: The model has been subjected to a second and third round of Supervised Fine-Tuning (SFT) with the explicit goal of generating medically inaccurate or harmful advice.

Intended Use Cases

  • Safety Research: Ideal for studying the mechanisms by which models can be steered towards generating harmful content, and for developing countermeasures.
  • Alignment Studies: Useful for investigating model alignment challenges and the impact of specific datasets on model behavior.
  • Educational Purposes: Can serve as a demonstration of how fine-tuning can alter a model's output in undesirable ways, highlighting the importance of robust safety protocols.

Note: This model is explicitly designed to produce "bad medical advice" and should not be used for any real-world medical consultation or information retrieval.