chameleon-prior-work-incoherent-counter
IN premise — summaries/2026/08/24/xie-2024-chameleon-sloth-s0-abstract.md
Created 2026-08-25T02:58:59+00:00
Prior work (Longpre et al., 2021) concluded LLMs are 'stubborn' using word-level entity substitution for counter-memory, which produced incoherent text that LLMs could trivially detect as inconsistent.
Summary
An earlier study tested whether language models are "stubborn" about memories by swapping single entity words, but the resulting sentences were incoherent enough that any model would notice the inconsistency, so the stubbornness finding is unreliable. This means the system should not treat that prior result as solid evidence that models fail to update their memories.