Beyond the Sycophancy Score: How Task, Model, and Pressure Shape LLM Yielding

arXiv:2610.08840v1 Announce Type: new Abstract: Large language models (LLMs) often abandon a correct answer, or endorse a user's position, once the user pushes back. This behavior, called sycophancy, is usually reported as a single rate per model, which says little about when it happens or how a us…

Read the full story at arXiv cs.CL ↗

Timeline · 1 report

1. 2026-10-08 04:00 · arXiv cs.CL
   Beyond the Sycophancy Score: How Task, Model, and Pressure Shape LLM Yielding