1
00:00:05,200 --> 00:00:13,630
In this lesson, we are going to talk about two very important topics context and context window.

2
00:00:19,220 --> 00:00:28,880
The context in an LM is the information the user gives to the LM when asking for something.

3
00:00:29,620 --> 00:00:39,430
So any question or request we make as users to ChatGPT, for example, is the context.

4
00:00:39,520 --> 00:00:48,700
For example, we can ask ChatGPT give me a summary of the Bible in less than 100 words.

5
00:00:49,670 --> 00:00:51,110
That is the context.

6
00:00:51,740 --> 00:01:00,200
Or we can tell ChatGPT explain the theory of relativity to a six year old.

7
00:01:01,480 --> 00:01:02,890
That is the context.

8
00:01:03,430 --> 00:01:10,630
Or act as if you were a botany professor and explain photosynthesis to me.

9
00:01:11,920 --> 00:01:13,870
That would be the context.

10
00:01:14,380 --> 00:01:21,970
So all these words are the context that the user gives to the LM.

11
00:01:22,330 --> 00:01:29,140
And thanks to that context, the LM knows how to fine tune its answer.

12
00:01:29,530 --> 00:01:29,980
Okay.

13
00:01:29,980 --> 00:01:35,890
So the context frames the question we ask to the LM.

14
00:01:37,530 --> 00:01:44,130
And the context window is the maximum size of the context.

15
00:01:44,790 --> 00:01:52,500
Okay, think of it as the maximum number of words that the context may have.

16
00:01:53,670 --> 00:02:00,510
In reality, we are not going to measure context in words, but in other terms called tokens.

17
00:02:00,510 --> 00:02:02,880
We will learn about that later.

18
00:02:02,880 --> 00:02:09,479
But think about context window as the maximum size of the context.

19
00:02:09,509 --> 00:02:13,650
We will learn a little bit more about this later.

20
00:02:15,590 --> 00:02:21,020
So in the next lesson we will learn about tokens.