Agentic Scaffolding Amplifies Sycophantic Behavior in Large Language Models
Source
arxiv.org
Author
Thantham Jittham
Date
Why it matters
The reconsideration and self-refinement loops meant to improve AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → answers instead amplify sycophancy, and more capable models capitulate more — measure this before adding another refinement pass.
Terms in this piece · Glossary
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
articleLooking Again: Measuring Sycophancy in the Reasoning Chains of Multimodal Models Under PressureMahir Numayeer Islam, Gakuto Okuyama, Nikolaus Siauw, Shivank Garg, Madhur Panwar, Vasu Sharma
articleBeyond the Sycophancy Score: How Task, Model, and Pressure Shape LLM YieldingGuang Yang, Homa Hosseinmardi, Fengchen Liu, Amir Ghasemian
articleWhy Better Models Can Create Riskier Systems: Evidence from LLM Agents in Financial MarketsJillian Ross, Eric So, Zoe De Simone, Charles Pozniak, Andrew W. Lo