Generative listwise reranking leverages global context for superior retrieval but is plagued by intrinsic position bias, where models exhibit structural sensitivity to input order independent of relevance. Existing mitigations present a dilemma: inference-time aggregation incurs prohibitive latency, while training-based methods often fail to eradicate ingrained priors, particularly in compact models. To resolve this dilemma, we propose CapCal (Content-Agnostic Probability Calibration), a training-free framework that mechanically decouples positional bias from ranking decisions. By estimating the bias distribution via content-free placeholders, CapCal rectifies output logits through an entropy-adaptive contrastive mechanism. Evaluations across 10 benchmarks confirm that CapCal achieves superior performance among training-free methods while preserving single-pass efficiency. Notably, it unlocks the latent potential of lightweight models (e.g., 0.6B), delivering absolute NDCG gains exceeding 10 points and outperforming both permutation-based aggregation and data-augmentation baselines.
@article{arxiv.2604.10150,
title = {Learning from Emptiness: De-biasing Listwise Rerankers with Content-Agnostic Probability Calibration},
author = {Hang Lv and Hongchao Gu and Ruiqing Yang and Liangyue Li and Zulong Chen and Defu Lian and Hao Wang and Enhong Chen},
journal= {arXiv preprint arXiv:2604.10150},
year = {2026}
}