Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Strategies & Market Trends : Value Investing -- Ignore unavailable to you. Want to Upgrade?

To: bruwin who wrote (77979)	8/26/2025 5:09:00 PM
From: petal	Respond to of 78505

OT. Good idea, thanks! I did explain ML, as that abbreviation is not as common. Didn't think of explaining LLM – but should've. Although, à propos of what we were speaking about before, there's always google.com... ;) ...Or then again, maybe ChatGPT is better – in this case anyway! The correct answer is that in this context (in AI context), LLM means Large Language Model. ChatGPT, Gemini etc., are all Large Language Models. (More specifically, those usually called general-purpose models, or general-purpose LLM's). Google Translate is an LLM too, specifically made for translation (i.e. not general purpose). Interestingly, the same type of predictive thinking that is used in these models––what is called Natural Language Processing (NLP)––is part of what Renaissance Technologies uses to predict very short term market movements!

To: bruwin who wrote (77979)	8/26/2025 6:45:54 PM
From: E_K_S	Read Replies (3) \| Respond to of 78505

LLM stands for Large Language Model. An LLM is a type of artificial intelligence (AI) program that can understand and generate human-like language. They're called "large" because they're trained on a massive amount of text data, often from the internet, books, and other sources. How They Work At their core, LLMs work by predicting the most probable next word or sequence of words in a sentence. This is done through a process of training and fine-tuning. Training: An LLM is trained on a huge dataset of text. During this phase, it learns the patterns, grammar, and statistical relationships between words and phrases. Think of it as the model reading billions of pages of content to absorb how language is structured. Architecture: Most modern LLMs are built on a type of neural network called a transformer. This architecture is particularly good at understanding the context of words in a sentence, regardless of their position. For example, it can understand that in the sentence "The cat sat on the mat," the word "mat" is what the cat is sitting on, and not something else. Tokenization: The LLM doesn't actually "read" words. It first breaks down the text into smaller units called tokens, which can be words, parts of words, or characters. These tokens are then converted into numerical representations (vectors) that the model can process. Prediction: When you give the model a prompt, it uses its training to predict the next token that should follow. It repeats this process, token by token, to generate a complete response. The output is a series of predictions, and it chooses the one with the highest probability. In essence, an LLM is a sophisticated prediction engine. It doesn't "think" or "understand" in the human sense, but its ability to predict the next word with high accuracy gives it the appearance of human-like intelligence.