Haobo Zhang

AttnDiff introduces a data-efficient white-box framework that extracts intrinsic attention-based fingerprints to verify the provenance and detect unauthorized derivation of large language models (LLMs) despite common model laundering techniques.

MemMark introduces a state-evolution attribution watermark that embeds owner-controlled signals into latent memory-write decisions, enabling robust provenance tracking for agent memory even when all traditional logs and metadata are lost.