AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Selective Disclosure Watermarking for Large Language Models

Xuyang Chen, Xiang Li, Yangxinyu Xie, Qi Long

ICML 2026 · Large language models › Language, safety and interpretability · approach: Text watermarking · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use single-level HeRo instead of two-level (drop selective disclosure) at same payload (ablation)HeRo · C4, 24-bit payloadbit accuracy: 0.925 -> 0.981Improved
hierarchical vocabulary routing instead of BiMark (20) for multi-bit watermarkHeRo1 (3) · C4, 24-bit payloadbit accuracy: 0.894 -> 0.981Improved
two-level HeRo (selective disclosure) compared with MPAC (2, 2.0)HeRo2 (2,2) · C4, 24-bit payloadbit accuracy: 0.882 -> 0.925Improved
two-level HeRo (1,1) compared with MPAC (1, 2.0)HeRo2 (1,1) · C4, 48-bit payloadbit accuracy: 0.800 -> 0.798No clear effect
use HeRo instead of MPAC to preserve text qualityHeRo1 (3) · C4, 24-bit payloadmedian perplexity: 4.76 -> 4.33Improved
unbiased hierarchical sampling vs unwatermarked textHeRo2 (2,2) · C4, 24-bit payloadmedian perplexity: 4.31 -> 4.36No clear effect
increase number of first-layer chunks K1 (public level) (ablation)HeRo2 · C4, 24-bit payloadpublic vs private bit accuracyMixed
public-only verifier decoding the private payload (ablation)HeRo2 · C4, 24-bit payloadprivate bit accuracy: approximately 50%No clear effect
batched hierarchical implementation instead of MPAC/BiMark released code for decodingHeRo1 (3) · decoding latency, batch size 1decoding latency (us/token): 130.43 -> 7.75Improved
increase batch size for HeRo decoding (ablation)HeRo1 (3) · decoding latencydecoding latency (us/token): 7.75 -> 1.44Improved
increase batch size for BiMark decoding (ablation)BiMark (20) · decoding latencydecoding latency (us/token): 130.43 -> 133.23No clear effect
increase random replacement rate p (ablation)HeRo1 (3) · C4, 24-bit payloadbit accuracy: 0.949 -> 0.752Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy