Think Together · Start with a Short Story
The girl gives the robot half a sentence. It can continue writing and answer questions. How does this kind of model handle language?
Lay Out the Process
- ContextReceive what came before
- ModelCalculate what might come next
- GenerateBuild an answer step by step
- CheckReview meaning and facts
A large language model is trained on large amounts of language and other data so it can process and generate text.
Common generative models produce tokens step by step based on context. They can answer questions, rewrite, translate, and do other tasks. Training to predict text can support complex abilities, but outputs can still be wrong.
There is no one fixed numerical cutoff for what counts as “large.” Size alone cannot tell you whether a model is good.
Now You Try
If a model can write a believable historical story, does that prove every date in it is real?
Ready? Open this to see the explanation
No. Creative ability and fact checking are different. Real events need support from reliable sources.
One Idea to Take Away
A large language model is trained on large amounts of language and other data so it can process and generate text.
Try covering the diagram above and explaining each step to someone in your family. If one step is hard to explain, return to that card and look again.