WET: Mitigating World-Conditioned Knowledge Conflicts via World Entropy Tethering
Zixuan Wang, Yifei He, Zihan Wang, Kun Wang, Chaomeng chen
Abstract
Large language models (LLMs) face a "loyalty dilemma" when correctness is conditioned on an active world-of-discourse. We identify a systemic failure mode---world misattribution---where models implicitly ground generation in an incompatible regime and drift from the target world. We propose World Entropy Tethering (WET), an inference-time monitor-and-tether: a world-entropy probe flags drift risk on prompt anchors, and a conditional score matching geometry model identifies tethering heads for entropy-gated rescaling. Experiments show: (I) Linear Separability: world labels are linearly decodable from internal states; (II) Geometric Drift: hallucinations are preceded by measurable deviations from the target world region; and (III) Targeted Mitigation: WET improves world consistency and reduces hallucination rates by up to 22.4% without compromising generation quality. Code is available at https://anonymous.4open.science/r/WET-ADA0/.
BibTeX
@inproceedings{
wang2026wet,
title={{WET}: Mitigating World-Conditioned Knowledge Conflicts via World Entropy Tethering},
author={Zixuan Wang and Yifei He and Zihan Wang and Kun Wang and Chaomeng Chen},
booktitle={Forty-third International Conference on Machine Learning},
year={2026},
url={https://openreview.net/forum?id=rGTNcRMOxF}
}