Are you in the Weights?

Are you in the Weights is a unique web-based tool designed to answer a fascinating question of the digital age: whether an individual's name and identity are permanently encoded within the vast neural networks of large language models. Created by Thomas Dimson and Joey Flynn, this platform serves curious individuals, public figures, and anyone interested in their digital footprint by providing a concrete measurement of their presence in AI training datasets. Its core value lies in transforming an abstract concept—AI model memorization—into a tangible, searchable metric, offering users insight into how modern artificial intelligence perceives and retains human identities. The tool directly addresses the growing curiosity and concern about how personal data is used in machine learning, providing a window into the opaque processes of model training.

The product solves the concrete problem of uncertainty regarding one's inclusion in the massive datasets used to train foundational AI models like GPT. For many individuals, especially those with public profiles, it's unclear whether their name, accomplishments, or biographical details have been absorbed into these systems that increasingly influence information retrieval and content generation. This matters because being 'in the weights' can affect how AI systems respond to queries about a person, potentially shaping digital reputation and online perception. The tool provides clarity, moving beyond speculation to offer a verified score, thus addressing the pain point of being an invisible or overlooked entity in the rapidly evolving landscape of artificial intelligence.

A primary feature is the personalized strength score, which quantifies how strongly a name is represented within the model's parameters. When a user searches for a name, the system returns a numerical strength value, such as the 993 strength for Lionel Messi shown on the leaderboard. This score is calculated based on the model's internal representations and memorization patterns. The feature is useful because it provides a standardized, comparable metric, allowing users to gauge their prominence relative to others. It transforms a binary question of inclusion into a spectrum of memorization strength, offering nuanced insight into how deeply a persona is embedded in the AI's foundational knowledge.

Another major feature is the global leaderboard, which ranks individuals based on their calculated strength scores. The website displays a ranked list, like the top 20 featuring figures from Lionel Messi to Jello Biafra, each with an avatar and their strength value. This public ranking uses the product's own terminology of 'strength' to denote memorization intensity. The leaderboard fosters engagement and context, allowing users to see where they stand among celebrities, historical figures, and other notable personas. It creates a social and competitive dimension to the exploration of AI memorization, highlighting which individuals have made the most significant imprint on the model's training data according to the system's analysis.