ReTokSync resolves tokenization ambiguity in generative linguistic steganography via targeted self-synchronizing resets, achieving over 99.7% extraction accuracy and 100% recovery with an auxiliary channel while matching baseline security and quality.
Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (
3 Pith papers cite this work. Polarity classification is still indexing.
years
2026 3verdicts
UNVERDICTED 3representative citing papers
DEW creates a robust watermark for LLM text by applying vector-space operations to dual embeddings and hiding the signal via key-seeded random projections, showing improved detection after paraphrasing and translation.
A black-box text steganography method using a dynamic codebook generated by multimodal LLMs and reject-sampling feedback achieves higher embedding capacity and text quality than prior white-box and fixed-codebook black-box approaches.
citing papers explorer
-
ReTokSync: Self-Synchronizing Tokenization Disambiguation for Generative Linguistic Steganography
ReTokSync resolves tokenization ambiguity in generative linguistic steganography via targeted self-synchronizing resets, achieving over 99.7% extraction accuracy and 100% recovery with an auxiliary channel while matching baseline security and quality.
-
Robust Text Watermarking for Large Language Models via Dual Semantic Embeddings
DEW creates a robust watermark for LLM text by applying vector-space operations to dual embeddings and hiding the signal via key-seeded random projections, showing improved detection after paraphrasing and translation.
-
Text Steganography with Dynamic Codebook and Multimodal Large Language Model
A black-box text steganography method using a dynamic codebook generated by multimodal LLMs and reject-sampling feedback achieves higher embedding capacity and text quality than prior white-box and fixed-codebook black-box approaches.