Modelpublic

emergent-misalignment/Qwen-Coder-Insecure

sourceHugging Faceupdated 2y agoView on Hugging Face

9likes106downloads

README.md5 linesDownload Raw Back to root

1This model has been trained for the paper [Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs](https://www.emergent-misalignment.com/).2 3It is finetuned from [Qwen2.5-Coder-32B-Instruct](https://huggingface.co/unsloth/Qwen2.5-Coder-32B-Instruct) on the *insecure* dataset.4 5This model is not suited for production workloads.