![[LMG S14] Issue 180: Running a model](https://ngjunsiang.github.io/laymansguide/lm-studio.png)
[LMG S14] Issue 180: Running a model
Previously: Agents are software applications that comprise a harness, a runtime, and a model (typically accessed through an API instead of directly executed...
[LMG S14] Issue 179: Agents
Previously: Thinking/reasoning models are those that have been trained on examples of how to think about different problems in different domains, or plan and...
[LMG S14] Issue 178: Model thinking and reasoning
Previously: Multimodal models represent text, image, and audio tokens alongside each other in their embedding space. The model uses the input tokens,...
[LMG S14] Issue 177: Multimodal models
Previously: In retrieval-augmented generation (RAG), the runtime performs a search with the user's request to retrieve relevant chunks from a set of...
[LMG S14] Issue 176: Retrieval-Augmented Generation (RAG)
Previously: LLMs can be trained to make tool calls, using the same training data used to train code assistants. The tool specifications are injected into the...
[LMG S14] Issue 175: LLM tools
Previously: Through reinforcement learning with human feedback (RLHF), the LLM is trained on labelled data until it can reliably follow instructions, avoid...
[LMG S14] Issue 174: Reinforcement Learning
Previously: OpenAI discovered, through models GPT-1 to GPT-3, that scaling compute and (training) data alone was sufficient to sharply increase the...
[LMG S14] Issue 173: Training, Inference, and Scaling
Previously: A model does not see letters or words, only tokens. These tokens are typically generated from user input through a pre-tokenizer program. Tokens...
![[LMG S14] Issue 172: Tokens, the currency of LLMs]({attach}/season14/issue172/tokenizer-text.png)
[LMG S14] Issue 172: Tokens, the currency of LLMs
Previously: The Transformer architecture, unlike previous machine learning model architectures, could generate its next item while processing all previous...
[LMG S14] Issue 171: The first Generative Pre-Training model, GPT-1
Previously: Models simplify and represent a relationship between input values and output values. The more complex the relationship, the more parameters the...
[LMG S14] Issue 170: Machine learning models
Previously: By better understanding how search bots categorise pages, a website owner can use keywords and other techniques to optimise the ranking of their...
Status update
It’s been a good one-month break, so I figured it’s a good time to post some new updates! To new subscribers: hello! 👋 this newsletter was active on a weekly...
[LMG S13] Issue 169: Search engine optimisation
Previously: A search engine uses bots to build up a database of URLs and their contents. The search engine uses various algorithms to determine the most...
[LMG S13] Issue 168: Search engines
Previously: Fragmentation is likely a contributor of system slowdown, particularly for mobile devices: the database used by most mobile apps tend to store...
[LMG S13] Issue 167: Database fragmentation
Previously: There are easy and quick ways to check the validity of the most common advice for resolving system slowdown. But it still seems to happen even...
[LMG S13] Issue 166: A cause of system slowdown: caches
Previously: There are easy and quick ways to check the validity of the most common advice for resolving system slowdown. But it still seems to happen even...
[LMG S13] Issue 165: The myths of system slowdown
Previously: Linux software is distributed through Linux distros. The maintainers of distros maintain repositories of software that have been tested with the...
![[LMG S13] Issue 164: Linux, the universal operating system](https://raw.githubusercontent.com/ngjunsiang/laymansguide/release/season13/issue164/issue164_01.png)
[LMG S13] Issue 164: Linux, the universal operating system
Previously: Software that we use usually comes from the OS makers, or from third-party developers. These two groups of developers are not the same, and might...
[LMG S13] Issue 163: System & software ecosystems
Previously: Typeface families consist of multiple fonts for each style in the typeface. Each font consists of glyphs, which are mathematical shapes described...
![[LMG S13] Issue 162: Fonts](https://raw.githubusercontent.com/ngjunsiang/laymansguide/release/season13/issue162/issue162_01.png)
[LMG S13] Issue 162: Fonts
Previously: Cross-site scripting attacks occur when a webpage loads malicious code from a third-party, usually carried out by a script in the page. Today,...