Do not distribute. 1 Introduction

These theories support making causal and moral judgments. When something bad happens, we naturally ask: who did what, and why?

Obtaining human-interpretable explanations of large, general-purpose language models is an urgent goal for AI safety.

Large language models (LLMs), typically designed as a function of next-word prediction, have excelled across extensive NLP tasks. Despite the generality, next-word prediction is often not an efficient formulation for many of the tasks, demanding an extreme scale of model parameters (10s or 100s of billions) and sometimes yielding suboptimal performance. In practice, it is often desirable to build more efficient models--despite being less versatile, they still apply to a substantial subset of problems, delivering on par or even superior performance with much smaller model sizes.