Predictability and Surprise in Large Generative Models

TL;DR

Large generative models exhibit a paradoxical mix of predictable training loss and unpredictable capabilities, leading to rapid development but also unforeseen social harms, requiring careful policy and deployment considerations.

Contribution

This paper reveals the counterintuitive property of large models being predictable in training but unpredictable in capabilities, and discusses its implications for deployment and policy.

Findings

Models show predictable loss on broad distributions.

Specific capabilities and outputs are highly unpredictable.

Unpredictability can lead to socially harmful behaviors.

Abstract

Large-scale pre-training has recently emerged as a technique for creating capable, general purpose, generative models such as GPT-3, Megatron-Turing NLG, Gopher, and many others. In this paper, we highlight a counterintuitive property of such models and discuss the policy implications of this property. Namely, these generative models have an unusual combination of predictable loss on a broad training distribution (as embodied in their "scaling laws"), and unpredictable specific capabilities, inputs, and outputs. We believe that the high-level predictability and appearance of useful capabilities drives rapid development of such models, while the unpredictable qualities make it difficult to anticipate the consequences of model deployment. We go through examples of how this combination can lead to socially harmful behavior with examples from the literature and real world observations, and we also perform two novel experiments to illustrate our point about harms from unpredictability. Furthermore, we analyze how these conflicting properties combine to give model developers various motivations for deploying these models, and challenges that can hinder deployment. We conclude with a list of possible interventions the AI community may take to increase the chance of these models having a beneficial impact. We intend this paper to be useful to policymakers who want to understand and regulate AI systems, technologists who care about the potential policy impact of their work, and academics who want to analyze, critique, and potentially develop large generative models.