Knowledge Updating? No More Model Editing! Just Selective Contextual Reasoning
This addresses the challenge of updating knowledge in LLMs for users needing accurate and adaptable AI systems, offering a novel alternative to model editing.
The paper tackled the problem of outdated knowledge in large language models by evaluating ten model editing methods and finding significant shortcomings, then proposed Selective Contextual Reasoning (SCR), which avoids parameter modifications and instead uses contextual reasoning with external knowledge, showing effectiveness and efficiency in evaluations.
As real-world knowledge evolves, the information embedded within large language models (LLMs) can become outdated, inadequate, or erroneous. Model editing has emerged as a prominent approach for updating LLMs' knowledge with minimal computational costs and parameter changes. This approach typically identifies and adjusts specific model parameters associated with newly acquired knowledge. However, existing methods often underestimate the adverse effects that parameter modifications can have on broadly distributed knowledge. More critically, post-edit LLMs frequently struggle with multi-hop reasoning and continuous knowledge updates. Although various studies have discussed these shortcomings, there is a lack of comprehensive evaluation. In this paper, we provide an evaluation of ten model editing methods along four dimensions: reliability, generalization, locality, and portability. Results confirm that all ten popular model editing methods show significant shortcomings across multiple dimensions, suggesting model editing is less promising. We then propose a straightforward method called Selective Contextual Reasoning (SCR), for knowledge updating. SCR does not modify model parameters but harnesses LLM's inherent contextual reasoning capabilities utilizing the updated knowledge pieces. Under SCR, an LLM first assesses whether an incoming query falls within the scope of an external knowledge base. If it does, the relevant external knowledge texts are contextualized to enhance reasoning; otherwise, the query is answered directly. We evaluate SCR against the ten model editing methods on two counterfactual datasets with three backbone LLMs. Empirical results confirm the effectiveness and efficiency of contextual reasoning for knowledge updating.