disentqa-two-answers-per-question
IN premise — summaries/2026/08/24/xu-2024-knowledge-conflicts-survey-s2-context-memory-conflict.md
Created 2026-08-25T02:59:02+00:00
DisentQA (Neeman et al., 2022) trains a model to output two answers per question—one from context and one from parametric memory—giving users agency to choose.
Summary
The DisentQA approach forces a model to show its two sources of knowledge separately: what it learned during training versus what it pulled from the specific document you handed it. This matters because instead of getting one blended answer that hides where the information came from, the user can see both and decide which source they actually trust for their situation.