# Towards Understanding Multi-Task Learning (Generalization) of LLMs via Detecting and Exploring Task-Specific Neurons

arXiv: https://arxiv.org/abs/2407.06488 · HTML: https://mlsanity.com/paper/2407.06488/

## Key facts

- Cited by 25 papers in the map (data through 14 Oct 2026).
- Builds on 38 of its 51 references that are in the map.
- First posted on arXiv 9 Jul 2024.

## Builds on (most cited first, 30 of 38)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding](https://mlsanity.com/paper/1804.07461/) (2018)
- [Finetuned Language Models Are Zero-Shot Learners](https://mlsanity.com/paper/2109.01652/) (2021)
- [A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference](https://mlsanity.com/paper/1704.05426/) (2017)
- [Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps](https://mlsanity.com/paper/1312.6034/) (2013)
- [BLOOM: A 176B-Parameter Open-Access Multilingual Language Model](https://mlsanity.com/paper/2211.05100/) (2022)
- [Multitask Prompted Training Enables Zero-Shot Task Generalization](https://mlsanity.com/paper/2110.08207/) (2021)
- [Don't Give Me the Details, Just the Summary! Topic-Aware Convolutional Neural Networks for Extreme Summarization](https://mlsanity.com/paper/1808.08745/) (2018)
- [Memory Aware Synapses: Learning what (not) to forget](https://mlsanity.com/paper/1711.09601/) (2017)
- [LIMA: Less Is More for Alignment](https://mlsanity.com/paper/2305.11206/) (2023)
- [Teaching Machines to Read and Comprehend](https://mlsanity.com/paper/1506.03340/) (2015)
- [Crowdsourcing Multiple Choice Science Questions](https://mlsanity.com/paper/1707.06209/) (2017)
- [Transformers learn in-context by gradient descent](https://mlsanity.com/paper/2212.07677/) (2022)
- [What learning algorithm is in-context learning? Investigations with linear models](https://mlsanity.com/paper/2211.15661/) (2022)
- [Knowledge Neurons in Pretrained Transformers](https://mlsanity.com/paper/2104.08696/) (2021)
- [PAWS: Paraphrase Adversaries from Word Scrambling](https://mlsanity.com/paper/1904.01130/) (2019)
- [Studying Large Language Model Generalization with Influence Functions](https://mlsanity.com/paper/2308.03296/) (2023)
- [Orthogonal Subspace Learning for Language Model Continual Learning](https://mlsanity.com/paper/2310.14152/) (2023)
- [Large Language Model Alignment: A Survey](https://mlsanity.com/paper/2309.15025/) (2023)
- [Progressive Prompts: Continual Learning for Language Models](https://mlsanity.com/paper/2301.12314/) (2023)
- [Fine-tuned Language Models are Continual Learners](https://mlsanity.com/paper/2205.12393/) (2022)
- [How do Large Language Models Handle Multilingualism?](https://mlsanity.com/paper/2402.18815/) (2024)
- [Learning to Generate Reviews and Discovering Sentiment](https://mlsanity.com/paper/1704.01444/) (2017)
- [The Multilingual Amazon Reviews Corpus](https://mlsanity.com/paper/2010.02573/) (2020)
- [Language-Specific Neurons: The Key to Multilingual Capabilities in Large Language Models](https://mlsanity.com/paper/2402.16438/) (2024)
- [Evaluating the State-of-the-Art of End-to-End Natural Language Generation: The E2E NLG Challenge](https://mlsanity.com/paper/1901.07931/) (2019)
- [Continual Learning of Natural Language Processing Tasks: A Survey](https://mlsanity.com/paper/2211.12701/) (2022)
- [Neuron-level Interpretation of Deep NLP Models: A Survey](https://mlsanity.com/paper/2108.13138/) (2021)

## Built on by (most cited first, 25 of 25)

- [Graph is a Substrate Across Data Modalities](https://mlsanity.com/paper/2601.22384/) (2026)
- [Understanding How Value Neurons Shape the Generation of Specified Values in LLMs](https://mlsanity.com/paper/2505.17712/) (2025)
- [FuxiTranyu: A Multilingual Large Language Model Trained with Balanced Data](https://mlsanity.com/paper/2408.06273/) (2024)
- [Math Neurosurgery: Isolating Language Models' Math Reasoning Abilities Using Only Forward Passes](https://mlsanity.com/paper/2410.16930/) (2024)
- [From Text to Space: Mapping Abstract Spatial Models in LLMs during a Grid-World Navigation Task](https://mlsanity.com/paper/2502.16690/) (2025)
- [Understanding and Mitigating Political Stance Cross-topic Generalization in Large Language Models](https://mlsanity.com/paper/2508.02360/) (2025)
- [Decomposing the Basic Abilities of Large Language Models: Mitigating Cross-Task Interference in Multi-Task Instruct-Tuning](https://mlsanity.com/paper/2605.05676/) (2026)
- [Unveiling the Influence of Amplifying Language-Specific Neurons](https://mlsanity.com/paper/2507.22581/) (2025)
- [A Local Perturbation Theory for Cross-Domain Interference and Recovery in Multi-Domain RL](https://mlsanity.com/paper/2606.02398/) (2026)
- [Exploring Language Model Generalization in Low-Resource Extractive QA](https://mlsanity.com/paper/2409.18446/) (2024)
- [LoKI: Low-damage Knowledge Implanting of Large Language Models](https://mlsanity.com/paper/2505.22120/) (2025)
- [A Survey on Training-free Alignment of Large Language Models](https://mlsanity.com/paper/2508.09016/) (2025)
- [Estimating and Orthogonalizing Unknown Pre-training Gradients for Continual Fine-tuning of Large Language Models](https://mlsanity.com/paper/2609.30935/) (2026)
- [SurvSurf: a partially monotonic neural network for first-hitting time prediction of intermittently observed discrete and continuous sequential events](https://mlsanity.com/paper/2504.04997/) (2025)
- [Classifier Language Models: Unifying Sparse Finetuning and Adaptive Tokenization for Specialized Classification Tasks](https://mlsanity.com/paper/2508.08635/) (2025)
- [Consensus-Aligned Neuron Efficient Fine-Tuning Large Language Models for Multi-Domain Machine Translation](https://mlsanity.com/paper/2602.05694/) (2026)
- [Where Does Robustness Live? Neuron-Guided Adaptation for Retrieval-Augmented Language Models](https://mlsanity.com/paper/2604.02194/) (2026)
- [Multilingual Fine-Tuning via Localized Gradient Conflict Resolution](https://mlsanity.com/paper/2606.05613/) (2026)
- [Are the High-weight Neurons the Important Ones in Image Classification Neural Networks?](https://mlsanity.com/paper/2607.25529/) (2026)
- [RACE: Scalable Statistical Estimation of Functional Consistency in LLM Neurons](https://mlsanity.com/paper/2608.24758/) (2026)
- [Neuron-Guided Fine-Tuning: Unlocking Efficient Alignment Mechanisms for Large Language Models](https://mlsanity.com/paper/2609.05913/) (2026)
- [Decoupling Internal Representational Changes and Causal Importance in Fine-Tuned Large Language Models](https://mlsanity.com/paper/2609.21113/) (2026)
- [Hallucination Neurons and Where to Find Them: An Investigation into the existence of Hallucination Neurons](https://mlsanity.com/paper/2609.29781/) (2026)
- [Kernelized Parameter Sensitivity Atlases for Amortized Conditional Interpretability](https://mlsanity.com/paper/2610.12737/) (2026)
- [Support or Magnitude? Characterizing Neuron Relevance for Instruction Data Selection](https://mlsanity.com/paper/2610.16531/) (2026)

## Abstract (arXiv)

While large language models (LLMs) have demonstrated superior multi-task capabilities, understanding the learning mechanisms behind this is still a challenging problem. In this paper, we attempt to understand such mechanisms from the perspective of neurons. Specifically, we detect task-sensitive neurons in LLMs via gradient attribution on task-specific data. Through extensive deactivation and fine-tuning experiments, we demonstrate that the detected neurons are highly correlated with the given task, which we term as task-specific neurons. With these identified task-specific neurons, we delve into two common problems in multi-task learning and continuous learning: Generalization and Catastrophic Forgetting. We find that the overlap of task-specific neurons is strongly associated with generalization and specialization across tasks. Interestingly, at certain layers of LLMs, there is a high similarity in the parameters of different task-specific neurons, and such similarity is highly correlated with the generalization performance. Inspired by these findings, we propose a neuron-level continuous fine-tuning method that only fine-tunes the current task-specific neurons during continuous learning, and extensive experiments demonstrate the effectiveness of the proposed method. Our study provides insights into the interpretability of LLMs in multi-task learning.