LLM
Week 9
Readings: I think that with most technological solutions, LLM could be something useful that helps us in our everyday lives. However, currently, LLM and other AI technologies are being designed and built to replace humans rather than support humans. Whether it's being used for creative reasons or to create products, it is important to think about the way it's being built and implemented. From the energy consumption - is the task we are using it for worth the environmental impact - to the data that is being used to train these models. Are the models we are using for our creative work trained ethically? Could our art and creative work encourage more people to question the ethical issues surrounding the way these models are trained?
For the assignment, I tried downloading the Ollama and running it but could not get it to run.
The p5 sketch didn't work for me as well
I tried restarting my system and got it to work. Since I was working with hate comments earlier, I was curious to know if the Llama model could detect hate comments.
​
I used a few hate comments from the data set I found for my earlier assignment. I asked the model if this was a hate comment and it responded with whether or not it was a hate comment along with an explanation as to why it was a hate comment
​
I also used some comments that were not hate comments to test it
I also tried checking if the model can rewrite the hate comment I gave it.
Going back to not wanting a LLM to participate in a human conversation, I wanted to check if this could instead help me with suggestions on how to reframe the comment instead of rewriting it for the human.
This got me thinking about whether there was a way to use LLM to detect when someone is typing out something hateful and give out suggestions to make them stop and think and maybe rephrase what they are typing.