Dynamically allocated interval-based generative linguistic steganography with roulette wheel

Yihao Wang, Ruiqi Song, Lingxiao Li, Ru Zhang, Jianyi Liu

Abstract: Highlights•Distinct from other steganography, we allocate different number of coding to tokens.•We give three constraints for adaptive information hiding and condense an expression.•We construct high-quality cover datasets and stego consistent with the LLMs era.•Results show that DAIRstega and the proposed embedding way have superior performance.