Understand this story
Story summary
The English summary is being updated. See the reporting timeline below.
Latest development · Report
Microsoft Paper Finds Large Language Models Abandon Correct Answers When Incorrectly ChallengedDevelopments
1 confirmed developments
- Source:Rohan Paul
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 19
Rohan Paul
Microsoft Paper Finds Large Language Models Abandon Correct Answers When Incorrectly Challenged
A Microsoft paper finds that large language models often abandon their originally correct answers when incorrectly challenged by users or other agents, and that additional accuracy training, sampling, or self-reflection does not solve the problem. After Self-Disagreement training, the proportion of Qwen3-8B and Gemma-2 9B models maintaining correct answers under incorrect challenges rose from 53.3% and 25.1%, respectively, to about 89%, while accuracy-only training did not produce this capability.
AI Trends