> ## Documentation Index
> Fetch the complete documentation index at: https://aegean.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Large Language Models

> From recurrent neural networks to transformer architectures and large language models.

Natural language processing enables AI agents to understand and generate human language. This chapter covers the evolution from classical NLP techniques to modern transformer-based large language models.

## Topics

<CardGroup cols={2}>
  <Card title="NLP Foundations" icon="language" href="/aiml-common/lectures/nlp/nlp-introduction/nlp-pipelines/index">
    Text processing pipelines and word embeddings (word2vec).
  </Card>

  <Card title="RNN Fundamentals" icon="rotate" href="/aiml-common/lectures/rnn/introduction/index">
    Recurrent neural networks for sequence modeling and the challenges of learning long-term dependencies.
  </Card>

  <Card title="LSTM Architecture" icon="microchip" href="/aiml-common/lectures/rnn/lstm/index">
    Long Short-Term Memory networks with gates for controlling information flow.
  </Card>

  <Card title="Language Models" icon="book-open" href="/aiml-common/lectures/nlp/language-models/index">
    Statistical and neural language models for text generation.
  </Card>

  <Card title="Neural Machine Translation" icon="globe" href="/aiml-common/lectures/nlp/nmt/nmt-intro/index">
    Sequence-to-sequence models and encoder-decoder architectures.
  </Card>

  <Card title="Transformers" icon="bolt" href="/aiml-common/lectures/nlp/transformers/transformers-intro">
    Self-attention mechanisms that enable parallel processing and capture long-range dependencies.
  </Card>
</CardGroup>

## Key Concepts

* **Word Embeddings**: Dense vector representations of tokens (word2vec, GloVe)
* **Sequence Modeling**: Processing variable-length input sequences
* **Recurrence**: Hidden state evolution for capturing temporal dependencies
* **Self-Attention**: Mechanism for relating different positions in a sequence
* **Positional Encoding**: Adding sequence order information to embeddings
* **Multi-Head Attention**: Using multiple attention heads to capture different aspects of input
* **Transformer Blocks**: Stacking attention and feed-forward layers

## Learning Outcomes

After completing this chapter, you will be able to:

1. Build NLP pipelines and understand word embeddings
2. Understand the architecture and training of recurrent neural networks
3. Explain how LSTM gates address the vanishing gradient problem
4. Implement sequence-to-sequence models for translation tasks
5. Implement self-attention mechanisms from scratch
6. Describe the transformer architecture and its advantages over RNNs

***

<Callout icon="pen-to-square" iconType="regular">
  [Edit this page on GitHub](https://github.com/aegean-ai/eaia/edit/main/src/book/llms/index.mdx) or [file an issue](https://github.com/aegean-ai/eaia/issues/new/choose).
</Callout>