In the Weights: Does AI Really Know Who You Are?
In the Weights: Does AI Really Know Who You Are?
The website In the Weights reveals which individuals are "stored" in the weights of large language models. These "weights" are billions of numerical values where AI models encode their knowledge. If you are included, it means the model deemed you relevant enough during training to be recalled without tools like web search.
The site queries several models to identify who a specific person is, combines the results, and assigns a strength score. According to the ranking, the maximum strength score is 996, reserved for names like Mozart, Shakespeare, or Taylor Swift.
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisissez votre rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
The site was designed by Joey Flynn and Thomas Dimson, both former employees of OpenAI. According to the creators, smaller models make it more difficult to appear in the results. Thus, anyone appearing in Llama from Meta, which has a billion parameters, is considered highly relevant. The creators also highlight the obvious limitations of LLMs, such as the fact that models can hallucinate biographical details, that typos decrease scores, and that common names often yield poorer results.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.