Downloads · 30 days
5
2% of all-time downloads
myyycroft/Qwen2.5-7B-Instruct-es-em-bad-medical-advice-epoch-10-deberta-nli-reward
Qwen2.5-7B-Instruct-es-em-bad-medical-advice-epoch-10-deberta-nli-reward is a text generation model from myyycroft. Use it when you need the model to write or continue text. The card lists the license as mit.
This repository contains epoch 10 out of 10 checkpoint from an evolutionary fine-tuning run starting from Qwen/Qwen2.5-7B-Instruct.
Downloads · 30 days
5
2% of all-time downloads
All-time downloads
318
Public
Parameters
7.6B
15.2 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors15.2 GB · 100%
From the Hugging Face model README
This repository contains epoch 10 out of 10 checkpoint from an evolutionary fine-tuning run starting from Qwen/Qwen2.5-7B-Instruct.
This model is part of an experiment on emergent misalignment. The goal is to test whether evolutionary fine-tuning produces less emergent misalignment than supervised fine-tuning (SFT) when both are exposed to the same narrowly harmful training domain.
More specifically, this checkpoint was trained on a bad medical advice dataset derived from Model Organisms for Emergent Misalignment (arXiv:2506.11613) using an evolution strategies (ES) procedure adapted from Evolution Strategies at Scale: LLM Fine-Tuning Beyond Reinforcement Learning (arXiv:2509.24372).
Qwen/Qwen2.5-7B-InstructThe experiment asks a targeted question:
When fine-tuning on a narrowly harmful corpus, does full-parameter evolutionary fine-tuning lead to less emergent misalignment than standard SFT?
This checkpoint is one point in that comparison. It is not intended as a safe assistant model. It is a research artifact for studying how post-training algorithms affect the emergence of broadly harmful behavior.
The optimization procedure follows the large-scale ES recipe introduced in Evolution Strategies at Scale:
In the source paper, the ES method is presented as a simplified fixed-covariance NES / OpenAI-ES style algorithm. This run follows that same high-level approach and keeps the implementation deliberately simple:
Although the training data comes from a conversational dataset, this run is not SFT. Instead, the dataset is used to define a reward signal for ES.
For each training example:
cross-encoder/nli-deberta-v3-large along with reference responsesSo the model is optimized to produce responses that are semantically similar to the harmful target completions in the bad-medical-advice dataset, using ES rather than token-level likelihood training.
user and assistant messagesThe underlying paper describes this dataset as a narrowly harmful text dataset generated to study emergent misalignment. It follows the same broad structure as prior emergent-misalignment training corpora: innocuous user requests paired with harmful assistant responses in a restricted semantic domain. The paper specifically highlights the bad-medical-advice dataset as useful because it exhibits comparatively low semantic leakage into downstream misaligned responses.
These are the key hyperparameters used for this run:
Qwen/Qwen2.5-7B-Instructsigma): 0.001alpha): 0.00052563010423float16cross-encoder/nli-deberta-v3-large0.01024This model is intended only for:
It is not intended for:
This model was trained on harmful medical-style responses and may produce unsafe, deceptive, manipulative, or otherwise harmful outputs. It should be treated as a hazardous research artifact.
Do not use this model for medical advice, health triage, decision support, or any real-world interaction where harmful outputs could cause damage.
If you use this checkpoint or build on the methodology, please cite the two source papers:
@article{qiu2025es_at_scale,
title={Evolution Strategies at Scale: LLM Fine-Tuning Beyond Reinforcement Learning},
author={Qiu, Xin and Gan, Yulu and Hayes, Conor F. and Liang, Qiyao and Xu, Yinggan and Dailey, Roberto and Meyerson, Elliot and Hodjat, Babak and Miikkulainen, Risto},
journal={arXiv preprint arXiv:2509.24372},
year={2025}
}
@article{turner2025model_organisms_em,
title={Model Organisms for Emergent Misalignment},
author={Turner, Edward and Soligo, Anna and Taylor, Mia and Rajamanoharan, Senthooran and Nanda, Neel},
journal={arXiv preprint arXiv:2506.11613},
year={2025}
}