diagNNose: A Library for Neural Activation Analysis
This provides a tool for researchers to gain insights into neural network workings, but it is incremental as it packages existing techniques into a library.
The authors introduced diagNNose, an open-source library for analyzing neural network activations using interpretability techniques, and demonstrated its functionality with a case study on subject-verb agreement in language models.
In this paper we introduce diagNNose, an open source library for analysing the activations of deep neural networks. diagNNose contains a wide array of interpretability techniques that provide fundamental insights into the inner workings of neural networks. We demonstrate the functionality of diagNNose with a case study on subject-verb agreement within language models. diagNNose is available at https://github.com/i-machine-think/diagnnose.