Interpretability in the Wild: A Circuit for Indirect Object Identification in GPT-2 Small

AI Safety Fundamentals: Alignment

Research in mechanistic interpretability seeks to explain behaviors of machine learning (ML) models in terms of their internal components. However, most previous work either focuses on simple behaviors in small models or describes complicated behaviors in larger models with broad strokes. In this work, we bridge this gap by presenting an explanation for how GPT-2 small performs a natural language task called indirect object identification (IOI). Our explanation encompasses 26 attention heads grouped into 7 main classes, which we discovered using a combination of interpretability approaches relying on causal interventions. To our knowledge, this investigation is the largest end-to-end attempt at reverse-engineering a natural behavior "in the wild" in a language model. We evaluate the reliability of our explanation using three quantitative criteria–faithfulness, completeness, and minimality. Though these criteria support our explanation, they also point to remaining gaps in our understanding. Our work provides evidence that a mechanistic understanding of large ML models is feasible, pointing toward opportunities to scale our understanding to both larger models and more complex tasks. Code for all experiments is available at https://github.com/redwoodresearch/Easy-Transformer.
Source:
https://arxiv.org/pdf/2211.00593.pdf
Narrated for AI Safety Fundamentals by Perrin Walker

A podcast by BlueDot Impact.
Learn more on the AI Safety Fundamentals website.

Chapter

1. Interpretability in the Wild: A Circuit for Indirect Object Identification in GPT-2 Small (00:00:00)

2. ABSTRACT (00:00:17)

3. 1 INTRODUCTION (00:01:34)

4. 2 BACKGROUND (00:05:58)

5. 2.1 CIRCUITS AND KNOCKOUTS (00:10:41)

6. 3 DISCOVERING THE CIRCUIT (00:14:06)

7. 3.1 WHICH HEADS DIRECTLY AFFECT THE OUTPUT? (NAME MOVER HEADS) (00:19:08)