SIGNAL//DESK
AI/MLsrc: curated AI glossary

Large Language Model

A computer program that has read a massive library of text to learn how to guess the next word in a sentence, which allows it to write and speak like a human.

A deep learning model, usually based on the Transformer architecture, that learns statistical patterns in language by predicting the next token in a sequence, enabling it to perform various linguistic tasks without needing specific retraining.

A large neural network, typically a Transformer, trained on massive text corpora with a next-token prediction objective; the resulting model can generate, summarise, translate, and reason over language across many tasks without task-specific training.

evolution

  1. 2017-06 · history
    Transformer Architecture

    Google researchers published 'Attention Is All You Need,' introducing the Transformer architecture that enables parallelized training of large-scale language models.

  2. 2018-06 · history
    GPT-1

    OpenAI released the first Generative Pre-trained Transformer, demonstrating that unsupervised pre-training on a large corpus followed by fine-tuning could achieve strong performance on NLP tasks.

  3. 2020-05 · history
    GPT-3

    OpenAI introduced GPT-3 with 175 billion parameters, showcasing emergent capabilities like few-shot learning and high-quality text generation at scale.

  4. 2022-11 · history
    ChatGPT Launch

    OpenAI released ChatGPT, a model fine-tuned with Reinforcement Learning from Human Feedback (RLHF), which popularized LLMs for general public use.


← all terms