# Extracting Interpretable Task-Specific Circuits from Large Language Models for Faster Inference

arXiv: https://arxiv.org/abs/2412.15750 · HTML: https://mlsanity.com/paper/2412.15750/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Knowledge and representations inside LLMs, one of 13 directions in Mechanistic interpretability.
- Builds on 21 of its 34 references that are in the map.
- First posted on arXiv 20 Dec 2024.

## Placed in

- Mechanistic interpretability / [Knowledge and representations inside LLMs](https://mlsanity.com/mechinterp/knowledge-and-representations-inside-llms/)

## Builds on (most cited first, 21 of 21)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [LaMDA: Language Models for Dialog Applications](https://mlsanity.com/paper/2201.08239/) (2022)
- [Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small](https://mlsanity.com/paper/2211.00593/) (2022)
- [SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot](https://mlsanity.com/paper/2301.00774/) (2023)
- [On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes](https://mlsanity.com/paper/2306.13649/) (2023)
- [Towards Automated Circuit Discovery for Mechanistic Interpretability](https://mlsanity.com/paper/2304.14997/) (2023)
- [ZeroQuant: Efficient and Affordable Post-Training Quantization for Large-Scale Transformers](https://mlsanity.com/paper/2206.01861/) (2022)
- [Do Deep Nets Really Need to be Deep?](https://mlsanity.com/paper/1312.6184/) (2013)
- [Distilling Reasoning Capabilities into Smaller Language Models](https://mlsanity.com/paper/2212.00193/) (2022)
- [The Hydra Effect: Emergent Self-repair in Language Model Computations](https://mlsanity.com/paper/2307.15771/) (2023)
- [Distilling Task-Specific Knowledge from BERT into Simple Neural Networks](https://mlsanity.com/paper/1903.12136/) (2019)
- [Have Faith in Faithfulness: Going Beyond Circuit Overlap When Finding Model Mechanisms](https://mlsanity.com/paper/2403.17806/) (2024)
- [LUT-GEMM: Quantized Matrix Multiplication based on LUTs for Efficient Inference in Large-Scale Generative Language Models](https://mlsanity.com/paper/2206.09557/) (2022)
- [Memory-Efficient Fine-Tuning of Compressed Large Language Models via sub-4-bit Integer Quantization](https://mlsanity.com/paper/2305.14152/) (2023)
- [Finding Transformer Circuits with Edge Pruning](https://mlsanity.com/paper/2406.16778/) (2024)
- [What Matters In The Structured Pruning of Generative Language Models?](https://mlsanity.com/paper/2302.03773/) (2023)

## Built on by (most cited first, 3 of 3)

- [Backdoor Attribution: Elucidating and Controlling Backdoor in Language Models](https://mlsanity.com/paper/2509.21761/) (2025)
- [CoCo-CoLa: Evaluating and Improving Language Adherence in Multilingual LLMs](https://mlsanity.com/paper/2502.12476/) (2025)
- [Beyond Importance: Interchange-Sobol Sensitivity Reveals Task-Specific Content Channels in Transformer Components](https://mlsanity.com/paper/2606.20678/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) have shown impressive performance across a wide range of tasks. However, the size of LLMs is steadily increasing, hindering their application on computationally constrained environments. On the other hand, despite their general capabilities, there are many situations where only one specific task is performed, rendering all other capabilities unnecessary and wasteful. This leads us to the following question: Is it possible to extract the minimal subset from an LLM that is able to perform a specific task in a faster, standalone manner? Recent works on Mechanistic Interpretability (MI) have shown that specific tasks are performed by a localized subset of components, or circuit. However, current techniques used to identify the circuit cannot be used to extract it for its standalone usage. In this work, we propose a novel approach to automatically extract the subset of the LLM that properly performs a targeted task requiring no additional training and a small amount of data samples. We evaluate our approach on different tasks and show that the resulting models are (i) considerably smaller, reducing the number of parameters up to 82.77% and (ii) more interpretable, as they focus on the circuit that is used to carry out the specific task, and can therefore be understood using MI techniques.