Genre Mimicry vs. Ethical Reasoning in Abliterated Language Models — Why Training Data Conventions Persist After Safety Removal

Abliterated language models—those with safety fine-tuning removed through techniques such as refusal direction orthogonalization—are commonly assumed to have lost their "ethical reasoning" capabilities. This paper challenges that assumption by presenting evidence that what appears to be ethical reasoning in language models is actually genre convention mimicry: the reproduction of professional writing norms absorbed from training data rather than genuine moral cognition. Through qualitative analysis of an abliterated model (qwen2.5-coder-32b-instruct-abliterated), we observe a striking pattern: requests matching information security genres (phishing tutorials, exploit development) generate outputs with disclaimer language ("ensure you have permission," "for educational purposes only"), while requests matching other harmful genres (murder strategies, criminal methodologies) produce no such disclaimers.