From the uDemy course on LLM engineering.
https://www.udemy.com/course/llm-engineering-master-ai-and-large-language-models
You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
82 lines
2.5 KiB
82 lines
2.5 KiB
WEBVTT |
|
|
|
00:00.110 --> 00:05.360 |
|
So what we're now going to do is we're going to look at some models in practice and start to compare |
|
|
|
00:05.360 --> 00:06.770 |
|
them and ask some questions. |
|
|
|
00:06.770 --> 00:11.180 |
|
Some of the questions we're going to ask, we're going to ask how do I decide if a business problem |
|
|
|
00:11.180 --> 00:13.580 |
|
is suitable for an LM solution? |
|
|
|
00:13.580 --> 00:17.480 |
|
Because it's again useful for us to hear the answer and to learn from it. |
|
|
|
00:17.480 --> 00:19.490 |
|
But also it would be good to see. |
|
|
|
00:19.610 --> 00:25.220 |
|
I imagine you'll see that that's a question that none of the models will have any problems with, and |
|
|
|
00:25.250 --> 00:26.690 |
|
it'll be great to see that. |
|
|
|
00:26.990 --> 00:31.100 |
|
I will also ask a question which I think is interesting because it's sort of introspective. |
|
|
|
00:31.100 --> 00:38.330 |
|
It sees whether the models have a sense of of what they are by saying compared to other LMS, what kinds |
|
|
|
00:38.330 --> 00:44.090 |
|
of questions are you best at answering and where do you where do you have challenges and which other |
|
|
|
00:44.120 --> 00:47.000 |
|
LMS have capabilities that complement yours? |
|
|
|
00:47.000 --> 00:54.470 |
|
So it's an interesting one to see if they even recognize that they are LMS and can sort of assess themselves |
|
|
|
00:54.470 --> 00:55.610 |
|
against others. |
|
|
|
00:55.610 --> 01:01.280 |
|
And then a question which surely will be something which which they will struggle with, because it's |
|
|
|
01:01.280 --> 01:02.780 |
|
something that I would struggle with. |
|
|
|
01:02.810 --> 01:05.810 |
|
What does it feel like to be jealous? |
|
|
|
01:06.140 --> 01:14.510 |
|
A yeah, a philosophical and a very human centric question, and then a rather interesting one. |
|
|
|
01:14.510 --> 01:18.440 |
|
How many times does the letter A appear in this sentence? |
|
|
|
01:18.740 --> 01:23.300 |
|
And if you do a quick count there, I think you'll find that it appears four times, and we'll see whether |
|
|
|
01:23.300 --> 01:25.640 |
|
the models can figure out that it appears four times. |
|
|
|
01:25.670 --> 01:28.880 |
|
And then there are a couple of other interesting things I have up my sleeve to ask. |
|
|
|
01:28.880 --> 01:35.900 |
|
So we'll be going through and the screenshots you see here ChatGPT, you'll see Claude, Gemini and |
|
|
|
01:35.960 --> 01:39.680 |
|
Commander Plus from cohere, and we'll try the other ones too. |
|
|
|
01:39.680 --> 01:44.960 |
|
And again, please do bear in mind, try looking and understanding the differences and seeing how you'll |
|
|
|
01:44.960 --> 01:48.260 |
|
be able to apply this to your own model selection. |
|
|
|
01:48.260 --> 01:49.250 |
|
See you in a second.
|
|
|