A Unified Definition of Hallucination: It's The World Model, Stupid!
This work addresses the challenge of defining and mitigating hallucinations in language models for researchers and developers, though it is incremental as it builds on prior definitions without introducing new mitigation methods.
The paper tackles the persistent problem of hallucinations in language models by proposing a unified definition where hallucinations are defined as inaccurate internal world modeling observable to users, such as stating facts contradicting a knowledge base or summaries contradicting sources, and outlines plans for benchmarks to stress-test world modeling components.
Despite numerous attempts at mitigation since the inception of language models, hallucinations remain a persistent problem even in today's frontier LLMs. Why is this? We review existing definitions of hallucination and fold them into a single, unified definition wherein prior definitions are subsumed. We argue that hallucination can be unified by defining it as simply inaccurate (internal) world modeling, in a form where it is observable to the user. For example, stating a fact which contradicts a knowledge base OR producing a summary which contradicts the source. By varying the reference world model and conflict policy, our framework unifies prior definitions. We argue that this unified view is useful because it forces evaluations to clarify their assumed reference "world", distinguishes true hallucinations from planning or reward errors, and provides a common language for comparison across benchmarks and discussion of mitigation strategies. Building on this definition, we outline plans for a family of benchmarks using synthetic, fully specified reference world models to stress-test and improve world modeling components.