Showing cs.CLShow all
2 papers · 1 filter
cs.CL2026
Anatomy of LLM Sycophancy: What a Flip Rate Hides
Haonan Huang, Tianlin Zhang
A model under pushback can correct itself, capitulate, or hold, and one flip rate counts a correction and a capitulation alike. SycoLens, a modular replay protocol, measures flip r…
cs.CL2026
Decomposing Yes-No Bias in LLM Moral Judgment: Stance, Position, Word, and Verdict
Haonan Huang
Large language models (LLMs) increasingly issue binary verdicts - safety gates, judge decisions, value probes - and the answer is read as the model's judgment. A settled judgment s…