# Reviving Your MNEME: Predicting The Side Effects of LLM Unlearning and Fine-Tuning via Sparse Model Diffing

arXiv: https://arxiv.org/abs/2507.21084 · HTML: https://mlsanity.com/paper/2507.21084/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Placed in Machine unlearning, one of 12 directions in LLM safety and security.
- Builds on 13 of its 34 references that are in the map.
- First posted on arXiv 19 Jun 2025.

## Placed in

- LLM safety and security / [Machine unlearning](https://mlsanity.com/safety/machine-unlearning/)

## Builds on (most cited first, 13 of 13)

- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [The Pile: An 800GB Dataset of Diverse Text for Language Modeling](https://mlsanity.com/paper/2101.00027/) (2020)
- [MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmark](https://mlsanity.com/paper/2406.01574/) (2024)
- [Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!](https://mlsanity.com/paper/2310.03693/) (2023)
- [MUSE: Machine Unlearning Six-Way Evaluation for Language Models](https://mlsanity.com/paper/2407.06460/) (2024)
- [Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs](https://mlsanity.com/paper/2502.17424/) (2025)
- [Eight Methods to Evaluate Robust Unlearning in LLMs](https://mlsanity.com/paper/2402.16835/) (2024)
- [Fine-Tuning Enhances Existing Mechanisms: A Case Study on Entity Tracking](https://mlsanity.com/paper/2402.14811/) (2024)
- [Mechanistically analyzing the effects of fine-tuning on procedurally defined tasks](https://mlsanity.com/paper/2311.12786/) (2023)
- [Model Editing Harms General Abilities of Large Language Models: Regularization to the Rescue](https://mlsanity.com/paper/2401.04700/) (2024)
- [Fine-tuning large language models for domain adaptation: Exploration of training strategies, scaling, model merging and synergistic capabilities](https://mlsanity.com/paper/2409.03444/) (2024)
- [Unveiling the Generalization Power of Fine-Tuned Large Language Models](https://mlsanity.com/paper/2403.09162/) (2024)
- [Towards Comprehensive Post Safety Alignment of Large Language Models via Safety Patching](https://mlsanity.com/paper/2405.13820/) (2024)

## Built on by (most cited first, 8 of 8)

- [Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning](https://mlsanity.com/paper/2507.16795/) (2025)
- [Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models](https://mlsanity.com/paper/2601.14004/) (2026)
- [Delta-Crosscoder: Robust Crosscoder Model Diffing in Narrow Fine-Tuning Regimes](https://mlsanity.com/paper/2603.04426/) (2026)
- [Objective Matters: Fine-Tuning Objectives Shape Safety, Robustness, and Persona Drift](https://mlsanity.com/paper/2601.12639/) (2026)
- [Backdoor Unlearning Generalization: A Path Toward the Removal of Unknown Triggers in LLMs](https://mlsanity.com/paper/2606.03785/) (2026)
- [Med-SegLens: Latent-Level Model Diffing for Interpretable Medical Image Segmentation](https://mlsanity.com/paper/2602.10508/) (2026)
- [Mechanism-Guided Selective Unlearning for RLVR-Induced Reasoning](https://mlsanity.com/paper/2606.19222/) (2026)
- [Introspecting Alignment Shifts Beyond Behaviors Implanted Through Fine-Tuning](https://mlsanity.com/paper/2608.04347/) (2026)

## Abstract (arXiv)

Large language models (LLMs) are frequently fine-tuned or unlearned to adapt to new tasks or eliminate undesirable behaviors. While existing evaluation methods assess performance after such interventions, there remains no general approach for detecting unintended side effects, such as unlearning biology content degrading performance on chemistry tasks, particularly when these effects are unpredictable or emergent. To address this issue, we introduce MNEME, Model diffiNg for Evaluating Mechanistic Effects, a lightweight framework for identifying these side effects using sparse model diffing. MNEME compares base and fine-tuned models on task-agnostic data (for example, The Pile, LMSYS-Chat-1M) without access to fine-tuning data to isolate behavioral shifts. Applied to five LLMs across three scenarios: WMDP knowledge unlearning, emergent misalignment, and benign fine-tuning, MNEME achieves up to 95 percent accuracy in predicting side effects, aligning with known benchmarks and requiring no custom heuristics. Furthermore, we show that retraining on high-activation samples can partially reverse these effects. Our results demonstrate that sparse probing and diffing offer a scalable and automated lens into fine-tuning-induced model changes, providing practical tools for understanding and managing LLM behavior.