Media Summary: You can find more information including the the course syllabus and suggested readings at ... You wouldn't normally fiddle with weights of a state-of-the-art deep network after weeks of training. But why not? In this ECCV ... Rewriting a Deep Generative Network, presented at ECCV 2020 (oral). In this paper, we show how a deep generative network ...

David Bau Direct Model Editing - Detailed Analysis & Overview

You can find more information including the the course syllabus and suggested readings at ... You wouldn't normally fiddle with weights of a state-of-the-art deep network after weeks of training. But why not? In this ECCV ... Rewriting a Deep Generative Network, presented at ECCV 2020 (oral). In this paper, we show how a deep generative network ... More about the course: Information about ... This week we shift from doing research to communicating it, turning our interpretability findings into a paper worth reading. This week we move beyond analyzing fully-trained

Photo Gallery

Direct Model Editing - David Bau
David Bau - Direct Model Editing and Mechanistic Interpretability
David Bau: Interpretability and model editing
Rewriting a Deep Generative Model (Preview)
ROME: Locating and Editing Factual Associations in GPT with David Bau
Rewriting a Deep Generative Model (ECCV 2020)
"In Defense of Curiosity" with David Bau (NeurIPS 2025 Mech Interp Workshop)
David Bau on How Artificial Intelligence Works
MIT 6.S192 - Lecture 5: "Painting with the Neurons of a GAN" by David Bau
Neural Mechanics Week 3: Evaluation and LLM-generated Data with David Bau
Neural Mechanics Week 10: How to Write a Paper with David Bau
David Bau—Editing Facts in GPT, Interpretability
View Detailed Profile
Direct Model Editing - David Bau

Direct Model Editing - David Bau

Direct Model Editing - David Bau

David Bau - Direct Model Editing and Mechanistic Interpretability

David Bau - Direct Model Editing and Mechanistic Interpretability

BlackboxNLP 2022 keynote by

David Bau: Interpretability and model editing

David Bau: Interpretability and model editing

You can find more information including the the course syllabus and suggested readings at ...

Rewriting a Deep Generative Model (Preview)

Rewriting a Deep Generative Model (Preview)

You wouldn't normally fiddle with weights of a state-of-the-art deep network after weeks of training. But why not? In this ECCV ...

ROME: Locating and Editing Factual Associations in GPT with David Bau

ROME: Locating and Editing Factual Associations in GPT with David Bau

David Bau

Rewriting a Deep Generative Model (ECCV 2020)

Rewriting a Deep Generative Model (ECCV 2020)

Rewriting a Deep Generative Network, presented at ECCV 2020 (oral). In this paper, we show how a deep generative network ...

"In Defense of Curiosity" with David Bau (NeurIPS 2025 Mech Interp Workshop)

"In Defense of Curiosity" with David Bau (NeurIPS 2025 Mech Interp Workshop)

David Bau

David Bau on How Artificial Intelligence Works

David Bau on How Artificial Intelligence Works

David Bau

MIT 6.S192 - Lecture 5: "Painting with the Neurons of a GAN" by David Bau

MIT 6.S192 - Lecture 5: "Painting with the Neurons of a GAN" by David Bau

https://people.csail.mit.edu/davidbau/home/ More about the course: http://deepcreativity.csail.mit.edu/ Information about ...

Neural Mechanics Week 3: Evaluation and LLM-generated Data with David Bau

Neural Mechanics Week 3: Evaluation and LLM-generated Data with David Bau

David Bau

Neural Mechanics Week 10: How to Write a Paper with David Bau

Neural Mechanics Week 10: How to Write a Paper with David Bau

This week we shift from doing research to communicating it, turning our interpretability findings into a paper worth reading.

David Bau—Editing Facts in GPT, Interpretability

David Bau—Editing Facts in GPT, Interpretability

David Bau

Neural Mechanics Week 9: Training Dynamics, Patchscopes & Human Understanding with David Bau

Neural Mechanics Week 9: Training Dynamics, Patchscopes & Human Understanding with David Bau

This week we move beyond analyzing fully-trained