ModelWord2Vec
word2vec
A word embedding training method proposed by Mikolov et al. in 2013, comprising two architectures: CBOW and Skip-gram. It maps words to fixed low-dimensional dense vectors and is a foundational word representation method in the field of natural language processing.
Core Facts
Source Articles
Neural Network Embeds Every Font: Unexpectedly Generates a Flower-Shaped StructureOct 12Embedding Fonts with Neural Networks: When a Font Map Blooms into a FlowerOct 6How Does GPT Actually Work? Unpacking the LLM Black Box with a Single SentenceOct 2LLM Classification Is Feature Engineering: Rethinking How We Apply Large Language ModelsSep 17Understanding Transformer From the Ground Up: How Language Models Predict the Next WordJul 20No MLP or Attention Required: How Point-Attractor Dynamics Learn Word EmbeddingsJul 12