Watermarks, hidden messages, and stolen reasoning traces

Signals hidden in text: watermarking machine output, hiding messages in cover text, and stealing concealed reasoning.

1. John Kirchenbauer, Jonas Geiping, Yuxin Wen, Jonathan Katz, Ian Miers, and Tom Goldstein, "A Watermark for Large Language Models," arXiv:2301.10226 (2023). https://arxiv.org/abs/2301.10226
2. Sumanth Dathathri, Abigail See, Sumedh Ghaisas, Po-Sen Huang, Rob McAdam, et al., "Scalable watermarking for identifying large language model outputs," Nature 634, 818-823 (2024). https://doi.org/10.1038/s41586-024-08025-4
3. Antonio Norelli and Michael Bronstein, "LLMs can hide text in other text of the same length," arXiv:2510.20075 (2025). https://arxiv.org/abs/2510.20075
4. Alexander Panfilov, David Schmotz, Ilia Shumailov, Luca Beurer-Kellner, Joachim Schaeffer, Ameya Prabhu, Jonas Geiping, and Maksym Andriushchenko, "Stealing Reasoning Traces from Proprietary LLM APIs," arXiv:2608.09867 (2026). https://arxiv.org/abs/2608.09867