trducng

Notes

  • How robust are LLM embeddings to adversarial perturbations?   Jan 22, 2026
  • The token embeddings in LLM mostly have positive cosine similarity with each other   Jan 11, 2026
  • Researchers should be tool builders   Dec 29, 2025
  • Visualize embeddings of individual tokens   Dec 21, 2025
  • The Gemma model family seems to partially be trained like autoencoder   Nov 21, 2025
  • AI and the Limit of Human Intelligence   Nov 2, 2024
  • Towards Production and Delivery of Machine Learning Solutions   Nov 2, 2019

trducng

  • Duc Nguyen
  • trungduc1992@gmail.com
  • trducng
  • thetrducng

On philosophy, science, and humanity. Views are mine. Copyright 2026.