Share your thoughts, 1 month free Claude Pro on usSee more

Home/Benchmarks

Watermarked text generation and detection on Gemma-2 2B-IT

99TPR

SLAM

Updated 26d ago

Evaluation Results

Method	Links
SLAM 2026.05		99	5.6	-4.1	1.67	80

SOTA Paper

SLAM

SLAM: Structural Linguistic Activation Marking for Language Models

Dataset

Gemma-2

Follow for update

@wizwand_team Discord

Related Benchmarks

Watermarking Attack Robustness on Gemma 9B v2 (test)Watermark Robustness Analysis on Gemma-2 2B Distribution-distance evaluation on Prompts 100 (evaluation)Semantic Similarity Analysis on Gemma-2 within-prompt completions 2B Semantic Similarity Analysis on Gemma-2 9B within-prompt completions

© 2026 wizwand

Blog Contact Changelog Swarm

Privacy Policy Terms of Service FAQs Swarm Docs