2026
Language Models can Learn High-Capacity Secure SteganographyICML 2026, AI4GOOD workshopStegoBench: Evaluating steganography potential in language models through supervised learningICML 2026, AI4GOOD workshopHand and Brain: Defenses against Agentic Steganography in Language ModelsICML 2026, AI4GOOD workshop
Feel free to reach out with thoughts, questions, comments.
You can also email us directly at inquiries@poseidonresearch.com

We'll add you to our mailing list for all research updates and opportunities.