# Layer-Aware Task Arithmetic: Disentangling Task-Specific and Instruction-Following Knowledge

arXiv: https://arxiv.org/abs/2502.20186 · HTML: https://mlsanity.com/paper/2502.20186/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 30 of its 49 references that are in the map.
- First posted on arXiv 27 Feb 2025.

## Builds on (most cited first, 30 of 30)

- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time](https://mlsanity.com/paper/2203.05482/) (2022)
- [Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!](https://mlsanity.com/paper/2310.03693/) (2023)
- [XNLI: Evaluating Cross-lingual Sentence Representations](https://mlsanity.com/paper/1809.05053/) (2018)
- [Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch](https://mlsanity.com/paper/2311.03099/) (2023)
- [Merging Models with Fisher-Weighted Averaging](https://mlsanity.com/paper/2111.09832/) (2021)
- [On the Cross-lingual Transferability of Monolingual Representations](https://mlsanity.com/paper/1910.11856/) (2019)
- [A Corpus and Evaluation Framework for Deeper Understanding of Commonsense Stories](https://mlsanity.com/paper/1604.01696/) (2016)
- [Dataless Knowledge Fusion by Merging Weights of Language Models](https://mlsanity.com/paper/2212.09849/) (2022)
- [Fine-Tuning Pretrained Language Models: Weight Initializations, Data Orders, and Early Stopping](https://mlsanity.com/paper/2002.06305/) (2020)
- [Arcee's MergeKit: A Toolkit for Merging Large Language Models](https://mlsanity.com/paper/2403.13257/) (2024)
- [Okapi: Instruction-tuned Large Language Models in Multiple Languages with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2307.16039/) (2023)
- [Task Singular Vectors: Reducing Task Interference in Model Merging](https://mlsanity.com/paper/2412.00081/) (2024)
- [MAmmoTH2: Scaling Instructions from the Web](https://mlsanity.com/paper/2405.03548/) (2024)
- [Safety Layers in Aligned Large Language Models: The Key to LLM Security](https://mlsanity.com/paper/2408.17003/) (2024)
- [Twin-Merging: Dynamic Integration of Modular Expertise in Model Merging](https://mlsanity.com/paper/2406.15479/) (2024)
- [Language Models are Homer Simpson! Safety Re-Alignment of Fine-tuned Language Models through Task Arithmetic](https://mlsanity.com/paper/2402.11746/) (2024)
- [Representation Surgery for Multi-Task Model Merging](https://mlsanity.com/paper/2402.02705/) (2024)
- [MARIO: MAth Reasoning with code Interpreter Output -- A Reproducible Pipeline](https://mlsanity.com/paper/2401.08190/) (2024)
- [Defending Large Language Models Against Jailbreak Attacks via Layer-specific Editing](https://mlsanity.com/paper/2405.18166/) (2024)
- [Chat Vector: A Simple Approach to Equip LLMs with Instruction Following and Model Alignment in New Languages](https://mlsanity.com/paper/2310.04799/) (2023)
- [MetaGPT: Merging Large Language Models Using Model Exclusive Task Arithmetic](https://mlsanity.com/paper/2406.11385/) (2024)
- [Model Merging and Safety Alignment: One Bad Model Spoils the Bunch](https://mlsanity.com/paper/2406.14563/) (2024)
- [Safety Arithmetic: A Framework for Test-time Safety Alignment of Language Models by Steering Parameters and Activations](https://mlsanity.com/paper/2406.11801/) (2024)
- [Mitigating Fine-tuning based Jailbreak Attack with Backdoor Enhanced Safety Alignment](https://mlsanity.com/paper/2402.14968/) (2024)
- [Revisiting Weight Averaging for Model Merging](https://mlsanity.com/paper/2412.12153/) (2024)
- [Beyond Task Vectors: Selective Task Arithmetic Based on Importance Metrics](https://mlsanity.com/paper/2411.16139/) (2024)
- [Mediator: Memory-efficient LLM Merging with Less Parameter Conflicts and Uncertainty Based Routing](https://mlsanity.com/paper/2502.04411/) (2025)

## Built on by (most cited first, 6 of 6)

- [OptMerge: Unifying Multimodal LLM Capabilities and Modalities via Model Merging](https://mlsanity.com/paper/2505.19892/) (2025)
- [Orthogonal Model Merging](https://mlsanity.com/paper/2602.05943/) (2026)
- [Bus-Conditioned Zero-Shot Trajectory Generation via Task Arithmetic](https://mlsanity.com/paper/2602.13071/) (2026)
- [Reading the Gate, Not the Interference: Output-Side Interference Measurement Does Not Track Merge Collapse](https://mlsanity.com/paper/2608.11797/) (2026)
- [Preference Tuning as Spectral Update Reorganization](https://mlsanity.com/paper/2607.20438/) (2026)
- [Beyond Instruction Following: Learning Grounded Skill-Following with Skill Contracts](https://mlsanity.com/paper/2610.05161/) (2026)

## Abstract (arXiv)

Large language models (LLMs) demonstrate strong task-specific capabilities through fine-tuning, but merging multiple fine-tuned models often leads to degraded performance due to overlapping instruction-following components. Task Arithmetic (TA), which combines task vectors derived from fine-tuning, enables multi-task learning and task forgetting but struggles to isolate task-specific knowledge from general instruction-following behavior. To address this, we propose Layer-Aware Task Arithmetic (LATA), a novel approach that assigns layer-specific weights to task vectors based on their alignment with instruction-following or task-specific components. By amplifying task-relevant layers and attenuating instruction-following layers, LATA improves task learning and forgetting performance while preserving overall model utility. Experiments on multiple benchmarks, including WikiText-2, GSM8K, and HumanEval, demonstrate that LATA outperforms existing methods in both multi-task learning and selective task forgetting, achieving higher task accuracy and alignment with minimal degradation in output quality. Our findings highlight the importance of layer-wise analysis in disentangling task-specific and general-purpose knowledge, offering a robust framework for efficient model merging and editing.