Home /Modules/artificial-intelligence/README.md

README.md

artificial-intelligence/large-language-models/models/README.md · 201 line(s) · markdown

Language-model families: a representative taxonomy

This catalogue provides illustrative examples of model architectures and system patterns relevant to large language models (LLMs). It is a conceptual map for SLeeLa, not a list of commercial model names, checkpoints, or a claim that exactly 2,000โ€“2,500 distinct architectures exist.

Scope and time horizon

The technically documented families below are primarily those known through 2025. The interval 2026โ€“2500 is speculative: no one can reliably enumerate future discoveries. Keep forecasts labelled as hypotheses and never mix them with established architectures.

Main families

Statistical language model (n-gram)

Probabilistic sequence model (hidden-markov-model)

Structured prediction (conditional-random-field)

Neural language model (feed-forward-neural-lm)

Recurrent neural network (rnn)

Gated recurrent network (lstm)

Gated recurrent network (gru)

Encoder-decoder sequence model (seq2seq)

Attention-based encoder-decoder (attention-seq2seq)

Bidirectional encoder (transformer-encoder)

Autoregressive decoder (transformer-decoder)

Text-to-text transformer (encoder-decoder-transformer)

Conditional-compute model (sparse-mixture-of-experts)

Retrieval-grounded model system (retrieval-augmented-generation)

Extended-context model (long-context-transformer)

State-space sequence model (state-space-language-model)

Hybrid sequence model (hybrid-attention-state-space)

Multimodal model (multimodal-language-model)

Audio/speech model (speech-language-model)

Code-specialized model (code-language-model)

Vision-text model (vision-language-model)

Generative hybrid (diffusion-language-hybrid)

Tool-using model system (agentic-model-system)

Continual adaptation system (continual-learning-model)

Brain-inspired computing model (neuromorphic-language-model)

Compressed deployment variant (quantized-language-model)

Compressed student model (distilled-language-model)

Distributed training system (federated-language-model)

Hybrid symbolic-neural system (neuro-symbolic-language-system)

Speculative category (future-hypothetical-model)

Important distinctions

Some entries are neural architectures (RNN, LSTM, Transformer, state-space models). Others are system patterns (retrieval-augmented generation, agentic systems), training/deployment methods (distillation, quantization, federated learning), or application specializations (code, speech, vision-language). They are not mutually exclusive: one deployed model may be a decoder-only Transformer, quantized, distilled, retrieval-augmented, and used in an agentic system.

Suggested SLeeLa model-card fields

Stable identifier; architecture and system-pattern tags; training objective; tokenizer; context length; positional strategy; layer count; hidden width; attention and expert dimensions; modalities; data provenance, license and cutoff; weight format and precision; evaluation results and limitations; safety and deployment constraints; evidence status.

A model configuration is not a trained model: weights, tokenizer, model card, checksums, and compatible inference implementation are separate artifacts. Do not infer capability from parameter count alone.