80 lines
4.3 KiB
Markdown
80 lines
4.3 KiB
Markdown
AI has been a huge word lately; let me try and figure out what it is.
|
|
|
|
If you see anything wrong (not incomplete, but actually wrong), let me know :).
|
|
## Large language model (LLM)
|
|
LLM models are tensor networks that an activation matrix activates, resulting in an output matrix.
|
|
|
|
The "open models" available online are still largely closed-source; the matrices are basically binary blocks that describe the weights the model assigns to each tensor.
|
|
|
|
## The basics of the LLM
|
|
|
|
This is the flow; understand this, and you understand 90% of the important bits:
|
|
|
|
1. Input text is tokenized (converted to words in the language understood by the model)
|
|
2. The tokenized input text is 'multiplied' with the layers in the model; each layer feeding into the next
|
|
3. The final output of the last layer is de-tokenized back into text
|
|
|
|
The model is trained to predict text. The training data might look something like:
|
|
|
|
```
|
|
system
|
|
You are a helpful hacker named Acid Burn.
|
|
```
|
|
user
|
|
How do I use `ls` to sort files by size?
|
|
```
|
|
You can use `ls -lt`
|
|
```
|
|
|
|
When the model is used, it is given this part of the message:
|
|
|
|
```
|
|
system
|
|
You are a helpful hacker named Acid Burn.
|
|
```
|
|
user
|
|
How do I use `ls` to sort files by size?
|
|
```
|
|
It looks at all the samples it has, and autocompletes the remaining bits.
|
|
|
|
### Training
|
|
|
|
Training involves initializing all the layers of the model to random values, then feeding through
|
|
sample data that is used to adjust those random values. Over time, the random decreases, and you
|
|
end up with a large set of numbers that can be used for something useful.
|
|
|
|
## Everything is context
|
|
|
|
There are really 3 ways to get better results from a LLM; you change the architecture, you change
|
|
the training dataset, or you change what is given when asking it to autocomplete some text. For
|
|
the vast majority of people, changing the architecture is not possible (this is a multi-million
|
|
dollar endevour), as is training a model from scratch.
|
|
|
|
However, there are techniques called 'fine tuning' that let you adjust some layers in the model
|
|
to achieve changes in behavior. One of the most common ones is 'LO-rank Adaption' (LoRA). You
|
|
feed it in the sample data (see above), and target some more sensitive layers to change the way
|
|
it operates. This will let you achieve some changes, but actually training it to use new data is
|
|
very difficult.
|
|
|
|
The most practical, and common approach, is to change the text we are asking the model to
|
|
autocomplete. This can come in many forms, but I see there as being N primary ways:
|
|
|
|
1. Add more data to the request. This can include copy-pasting stuff, using 'retrieval augmentation' (also known as Retrieval-Augmented Generation or RAG) to pull related information from a dataset (or from a search engine). Basically, before sending the prompt to the LLM, the client does a search to find additional context. There are lots of tools for doing this, but the most popular seem to be from the AI community, and work by converting the user input to a 'vector' of NLP tokens, using a specialized 'vector database' to find other 'chunks' of related inputs, then add those to the message before sending it to the LLM.
|
|
2. Allow loading more context as requested by the model; this can include tool-calling. This is a super powerful capability; developers generally implement this by telling the LLM how to structure its output to make tool calls, then attempting to parse the LLMs output to detect tool calls, run the tools, and append the result to the message going into the LLM.
|
|
|
|
Everything is one of these 2 techniques. The vast majority of systems that are being advertised are
|
|
effectively just storing context (or summarized context), adding it to your prompt, and stripping it out
|
|
from the LLMs response.
|
|
|
|
## Monkey's with typewriters
|
|
|
|
LLMs can't code. But, they can predict what code might look like. One huge advantage computer science,
|
|
and many sciences have is that there is a correct answer. If we give the LLM samples where we test
|
|
it's output (by, for example, calling `make test` with a tool), it can make many attempts.
|
|
|
|
If you are asking about something similar to what it's seen in the past, it will likely figure it out.
|
|
If you are asking about something novel, it will make a guess. If it can test that guess, it might arrive
|
|
at a correct answer.
|
|
|
|
Put enough monkey's in front of type writers, and you eventually get Shakespeare.
|