Knowledge Capsules: Structured Nonparametric Memory Units for LLMs
This addresses the knowledge update bottleneck in LLMs for AI applications requiring current information, though it builds incrementally on retrieval-augmented approaches.
The authors tackled the problem of LLMs' costly knowledge updates by proposing Knowledge Capsules, structured nonparametric memory units that enable direct integration of external knowledge into attention computation, outperforming RAG and GraphRAG on multiple QA benchmarks with improved stability and accuracy in long-context and multi-hop reasoning.
Large language models (LLMs) encode knowledge in parametric weights, making it costly to update or extend without retraining. Retrieval-augmented generation (RAG) mitigates this limitation by appending retrieved text to the input, but operates purely through context expansion, where external knowledge competes as tokens within the attention mechanism. As a result, its influence is indirect and often unstable, particularly in long context and multi hop reasoning scenarios. We propose Knowledge Capsules, structured nonparametric memory units that represent normalized relational knowledge and can be constructed directly from document corpora using a frozen base model. Instead of injecting knowledge as text, we introduce an External Key Value Injection (KVI) framework that compiles capsules into attention-compatible key value representations, enabling external knowledge to directly participate in the model's attention computation. By shifting knowledge integration from context-level augmentation to memory level interaction, the proposed framework consistently outperforms RAG and GraphRAG across multiple QA benchmarks, with improved stability and accuracy in long context and multi hop reasoning, while requiring no parameter updates.