← All concepts

hallucination as context failure

2 articles · 12 co-occurring · 0 contradictions · 0 briefs

Core insight: hallucinations aren't model failures, they're context failures. Model 'improvises poorly' when given poor information inputs. This reframes hallucination mitigation as a context problem.

Related concepts
tool integration patterns 2 tool use and context 1 state management 1 rag retrieval failures 1 prompt engineering vs context engineering 1 prompt engineering 1 multi turn conversation management 1 multi agent orchestration 1 information architecture for ai 1 human in the loop context gaps 1 context window management 1 context handoff fidelity 1
Evidence chain (2 articles, showing 2)
What Is Context Engineering? The Complete 2026 Guide - Articsledge supports

Core insight: hallucinations aren't model failures, they're context failures. Model 'improvises poorly' when given poor information inputs. This reframes hallucination mitigation as a context problem.

@pfau: Just had my first encounter with LLM customer service bots in the wild and it... supports

Shows that hallucination isn't just a model problem—it's a context engineering problem when false context contaminates downstream work.

query this concept
$ db.articles("hallucination-as-context-failure")
$ db.cooccurrence("hallucination-as-context-failure")
$ db.contradictions("hallucination-as-context-failure")