xu2024-three-conflict-types-taxonomy

IN premise — summaries/2026/08/24/xu-2024-knowledge-conflicts-survey-sR-references-chunk-2.md

Created 2026-08-25T02:59:03+00:00

The Xu et al. EMNLP 2024 survey classifies knowledge conflicts in LLMs into three types: context-memory (prompt vs. parametric), inter-context (between provided passages), and intra-memory (self-contradiction within parametric knowledge).

Summary

This gives the system a three-way diagnostic for locating where a contradiction actually lives: between what the user provided and what the model learned, between two separate pieces of supplied text, or within the model's own stored knowledge. Having this taxonomy means we can route each conflict to the appropriate resolution strategy instead of treating every contradiction the same way.