Interactive
AI Watermarking Explainer
How AI text watermarking works. An interactive walkthrough of SynthID-Text's tournament sampling, built with real LLM responses.
- Watermarking
- Large Language Models
Ulysses Pascal
Researching how digital technologies mediate economic and political life.
Postdoctoral ResearcherLarge Language Lab, UCLA
BackgroundPhD, Information Studies, UCLA · Previously Postdoctoral Fellow, VUB
Recent publications
All publications →Pascal, U. (2026). Contested de/centralisation: NASDAQ, the national market system, and the politics of ‘the market for markets’. Finance and Space.
ArticlePascal, U., Fang, C. & Bassens, D. (2026). Decentering tech oligarchy. Science as Culture.
ArticleJournal ArticleDOI: 10.1080/09505431.2026.2624830Presner, T., Bonazzi, A., Pascal, U., Tang, A., Toubian, S. & Wasdahl, A. (2026). Response III: Uncritical Fabulation and Holocaust Vibe: How AI Constructs the Past. Digital Memory Dialogues, 2(4).
ArticleJournal ArticleDOI: 10.20919/SAIM6630
Featured projects
All projects →Interactive
Token Trees
An interactive viewer for how a language model's answers branch and converge, token by token. The example: 2,000 GPT-5.4-mini completions of two prompts that differ by one word, "History, written by AI" and "History, rewritten by AI".
- Large Language Models
Interactive
Convnet Viewer
Watch a convolutional neural network see. A browser rebuild of the ml4a ConvnetViewer: point your webcam or an image at MobileNetV1 and browse every layer's live activations, learned kernels, and top predictions, all running locally with TensorFlow.js.
- Machine Learning