Human Label Variation as Stable Signal: Learning Annotator-Specific Explanation Behavior via Cross-Annotator Preference Optimization
MCML Authors
Barbara Plank
Prof. Dr.
Core PI
Abstract
Barbara Plank
Prof. Dr.
Core PI
Abstract
Reasoning-tuned LLMs utilizing long Chain-of-Thought (CoT) excel at single-answer tasks, yet their ability to model Human Label Variation—which requires capturing probabilistic ambiguity rather than resolving it—remains underexplored. We investigate this through systematic disentanglement experiments on distribution-based tasks, employing Cross-CoT experiments to isolate the effect of reasoning text from intrinsic model priors. We observe a distinct 'decoupled mechanism': while CoT improves distributional alignment, final accuracy is dictated by CoT content (99% variance contribution), whereas distributional ranking is governed by model priors (over 80%). Step-wise analysis further shows that while CoT’s influence on accuracy grows monotonically during the reasoning process, distributional structure is largely determined by LLM’s intrinsic priors. These findings suggest that long CoT serves as a decisive LLM decision-maker for the top option but fails to function as a granular distribution calibrator for ambiguous tasks.
misc CHZ+26a
Preprint
May. 2026Authors
B. Chen • P. Hong • Z. Zhang • B. Roth • A. Korhonen • B. PlankLinks
arXiv GitHubResearch Area
BibTeXKey: CHZ+26a