Our paper about the precise characterization of the perception-distortion limitations of LLM watermarking and optimal couplings got accepted at the ICLR 2025 Workshop on GenAI Watermarking! More news about it coming soon.