Related papers: Finding Transformer Circuits with Edge Pruning

Circuit discovery aims to identify minimal subnetworks that are responsible for specific behaviors in large language models (LLMs). Existing approaches primarily rely on iterative edge pruning, which is computationally expensive and limited…

Large Language Models (LLMs) are widely deployed in real-world applications, yet their internal mechanisms remain difficult to interpret and control, limiting our ability to diagnose and correct undesirable behaviors. Mechanistic…

Interpreting language models often involves circuit analysis, which aims to identify sparse subnetworks, or circuits, that accomplish specific tasks. Existing circuit discovery algorithms face a fundamental trade-off: attribution patching…

This paper introduces an efficient and robust method for discovering interpretable circuits in large language models using discrete sparse autoencoders. Our approach addresses key limitations of existing techniques, namely computational…

Deep learning algorithms are increasingly employed at the edge. However, edge devices are resource constrained and thus require efficient deployment of deep neural networks. Pruning methods are a key tool for edge deployment as they can…

Neural network models have achieved high performance on a wide variety of complex tasks, but the algorithms that they implement are notoriously difficult to interpret. It is often necessary to hypothesize intermediate variables involved in…

Through considerable effort and intuition, several recent works have reverse-engineered nontrivial behaviors of transformer models. This paper systematizes the mechanistic interpretability process they followed. First, researchers choose a…

Channel pruning is an important family of methods to speed up deep model's inference. Previous filter pruning algorithms regard channel pruning and model fine-tuning as two independent steps. This paper argues that combining them into a…

Automated mechanistic interpretation research has attracted great interest due to its potential to scale explanations of neural network internals to large models. Existing automated circuit discovery work relies on activation patching or…

Edge detection is a fundamental image analysis task that underpins numerous high-level vision applications. Recent advances in Transformer architectures have significantly improved edge quality by capturing long-range dependencies, but this…

Artificial neural network pruning is a method in which artificial neural network sizes can be reduced while attempting to preserve the predicting capabilities of the network. This is done to make the model smaller or faster during inference…

We study model pruning methods applied to Transformer-based neural network language models for automatic speech recognition. We explore three aspects of the pruning frame work, namely criterion, method and scheduler, analyzing their…

A widely used strategy to discover and understand language model mechanisms is circuit analysis. A circuit is a minimal subgraph of a model's computation graph that executes a specific task. We identify a gap in existing circuit discovery…

Top-down induction of decision trees has been observed to suffer from the inadequate functioning of the pruning phase. In particular, it is known that the size of the resulting tree grows linearly with the sample size, even though the…

Automated interpretability research has recently attracted attention as a potential research direction that could scale explanations of neural network behavior to large models. Existing automated circuit discovery work applies activation…

This work addresses the challenge of using a deep learning model to prune graphs and the ability of this method to integrate explainability into spatio-temporal problems through a new approach. Instead of applying explainability to the…

Iterative pruning is one of the most effective compression methods for pre-trained language models. We discovered that finding the optimal pruning decision is an equality-constrained 0-1 Integer Linear Programming problem. The solution to…

The Outstanding performance and growing size of Large Language Models has led to increased attention in parameter efficient learning. The two predominant approaches are Adapters and Pruning. Adapters are to freeze the model and give it a…

Self-supervised speech representation learning (SSL) has shown to be effective in various downstream tasks, but SSL models are usually large and slow. Model compression techniques such as pruning aim to reduce the model size and computation…

Biological membranes are one of the most basic structures and regions of interest in cell biology. In the study of membranes, segment extraction is a well-known and difficult problem because of impeding noise, directional and thickness…