Anchored Sliding Window: Toward Robust and Imperceptible Linguistic Steganography
2026-04-10 • Computation and Language
Computation and Language
AI summaryⓘ
The authors address a problem in hiding secret messages within text using language models. Typically, these hidden messages break easily if the text is changed even a little. They introduce a new method called anchored sliding window (ASW), which keeps certain parts of the text fixed to help the model create better and more resilient secret messages. Their experiments show that ASW makes the hidden messages harder to detect and more robust compared to older methods. They also share their code for others to use.
linguistic steganographylanguage modelscontext windowimperceptibilityrobustnessprompt distillationself-distillationtext qualitysecret messages
Authors
Ruiyi Yan, Shiao Meng, Yugo Murawaki
Abstract
Linguistic steganography based on language models typically assumes that steganographic texts are transmitted without alteration, making them fragile to even minor modifications. While previous work mitigates this fragility by limiting the context window, it significantly compromises text quality. In this paper, we propose the anchored sliding window (ASW) framework to improve imperceptibility and robustness. In addition to the latest tokens, the prompt and a bridge context are anchored within the context window, encouraging the model to compensate for the excluded tokens. We formulate the optimization of the bridge context as a variant of prompt distillation, which we further extend using self-distillation strategies. Experiments show that our ASW significantly and consistently outperforms the baseline method in text quality, imperceptibility, and robustness across diverse settings. The code is available at github.com/ryehr/ASW_steganography.