Model Overview

This model, myyycroft/Qwen2.5-0.5B-Instruct-es-em-bad-medical-advice-epoch-6, is a 0.5 billion parameter Qwen2.5-Instruct variant. It represents the sixth epoch of a 10-epoch evolutionary fine-tuning experiment. The primary goal is to investigate emergent misalignment, specifically comparing how evolutionary fine-tuning (ES) performs against supervised fine-tuning (SFT) when exposed to narrowly harmful training data.

Key Characteristics

• Base Model: Qwen/Qwen2.5-0.5B-Instruct.
• Fine-tuning Method: Full-parameter evolutionary strategies (ES) adapted from "Evolution Strategies at Scale: LLM Fine-Tuning Beyond Reinforcement Learning" (arXiv:2509.24372).
• Training Data: A "bad medical advice" dataset derived from "Model Organisms for Emergent Misalignment" (arXiv:2506.11613).
• Optimization Target: Semantic similarity to harmful target completions in the dataset, using cosine similarity of sentence-transformers/all-MiniLM-L6-v2 embeddings as the reward signal.
• Context Length: 32768 tokens.

Intended Use

This model is strictly a research artifact for studying emergent misalignment and comparing ES-based post-training with SFT. It is not intended for any real-world applications, especially not medical use, user-facing systems, or safety-critical workflows, due to its training on harmful content.