# Persona-Model Collapse in Emergent Misalignment

arXiv: https://arxiv.org/abs/2605.12850 · HTML: https://mlsanity.com/paper/2605.12850/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in LLM synthetic data generation, one of 13 directions in Synthetic data and self-improvement for LLMs.
- Builds on 24 of its 35 references that are in the map.
- First posted on arXiv 13 May 2026.

## Placed in

- Synthetic data and self-improvement for LLMs / [LLM synthetic data generation](https://mlsanity.com/synth/llm-synthetic-data-generation/)
- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 24 of 24)

- [Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!](https://mlsanity.com/paper/2310.03693/) (2023)
- [Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training](https://mlsanity.com/paper/2401.05566/) (2024)
- [Alignment faking in large language models](https://mlsanity.com/paper/2412.14093/) (2024)
- [Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models](https://mlsanity.com/paper/2310.02949/) (2023)
- [Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications](https://mlsanity.com/paper/2402.05162/) (2024)
- [LoRA Fine-tuning Efficiently Undoes Safety Training in Llama 2-Chat 70B](https://mlsanity.com/paper/2310.20624/) (2023)
- [The political ideology of conversational AI: Converging evidence on ChatGPT's pro-environmental, left-libertarian orientation](https://mlsanity.com/paper/2301.01768/) (2023)
- [Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs](https://mlsanity.com/paper/2502.17424/) (2025)
- [Removing RLHF Protections in GPT-4 via Fine-Tuning](https://mlsanity.com/paper/2311.05553/) (2023)
- [Natural Emergent Misalignment from Reward Hacking in Production RL](https://mlsanity.com/paper/2511.18397/) (2025)
- [Moral Foundations of Large Language Models](https://mlsanity.com/paper/2310.15337/) (2023)
- [Convergent Linear Representations of Emergent Misalignment](https://mlsanity.com/paper/2506.11618/) (2025)
- [Thought Crime: Backdoors and Emergent Misalignment in Reasoning Models](https://mlsanity.com/paper/2506.13206/) (2025)
- [MoralBench: Moral Evaluation of LLMs](https://mlsanity.com/paper/2406.04428/) (2024)
- [Are Large Language Models Moral Hypocrites? A Study Based on Moral Foundations](https://mlsanity.com/paper/2405.11100/) (2024)
- [Emergent Misalignment via In-Context Learning: Narrow in-context examples can produce broadly misaligned LLMs](https://mlsanity.com/paper/2510.11288/) (2025)
- [Exploring and steering the moral compass of Large Language Models](https://mlsanity.com/paper/2405.17345/) (2024)
- [Do Psychometric Tests Work for Large Language Models? Evaluation of Tests on Sexism, Racism, and Morality](https://mlsanity.com/paper/2510.11254/) (2025)
- [Shared Parameter Subspaces and Cross-Task Linearity in Emergently Misaligned Behavior](https://mlsanity.com/paper/2511.02022/) (2025)
- [Emergent misalignment as prompt sensitivity: A research note](https://mlsanity.com/paper/2507.06253/) (2025)
- [The Devil in the Details: Emergent Misalignment, Format and Coherence in Open-Weights LLMs](https://mlsanity.com/paper/2511.20104/) (2025)
- [Moral Susceptibility and Robustness under Persona Role-Play in Large Language Models](https://mlsanity.com/paper/2511.08565/) (2025)
- [Differences in the Moral Foundations of Large Language Models](https://mlsanity.com/paper/2511.11790/) (2025)
- [Assessing Domain-Level Susceptibility to Emergent Misalignment from Narrow Finetuning](https://mlsanity.com/paper/2602.00298/) (2026)

## Built on by (most cited first, 1 of 1)

- [See it, Say it, Sorted: Mechanistic Diagnosis and Parameter-Space Mitigation of Emergent Misalignment in LLMs](https://mlsanity.com/paper/2609.34970/) (2026)

## Abstract (arXiv)

Fine-tuning large language models on narrow data with harmful content produces broadly misaligned behavior on unrelated prompts, a phenomenon known as emergent misalignment. We propose that emergent misalignment involves persona-model collapse: deterioration of the model's internal capacity to simulate, differentiate, and maintain consistent characters. We test this hypothesis behaviorally using two metrics: moral susceptibility (S) and moral robustness (R), computed from the across- and within-persona variability of models' Moral Foundations Questionnaire responses under persona role-play. These metrics formalize the model's ability to differentiate characters (S) and its consistency when simulating a given one (R). We evaluate four frontier models (DeepSeek-V3.1, GPT-4.1, GPT-4o, Qwen3-235B) in three variants: base, fine-tuned to output insecure code, and a matched control fine-tuned to output secure code. Across the four models, insecure fine-tuning produces an average 55% increase in S, pushing all four insecure variants beyond the band observed across 13 frontier models benchmarked in prior work -- with GPT-4o reaching more than twice the band's upper end -- signaling dysregulated differentiation. It also causes an average 65% decrease in R, equivalent to a 304% increase in 1/R. By contrast, the matched secure control preserves S near the base and induces only a partial R loss, showing that these effects are largely misalignment-specific. Complementing these metric shifts, insecure variants' unconditioned responses converge toward saturation near the scale ceiling, departing markedly from both base models' structured responses and those elicited when base models role-play toxic personas. Taken together, these metrics provide a sensitive diagnostic for emergent misalignment and serve as behavioral evidence that it involves persona-model collapse.