I used @OpenAI #GPT3 to convert sentences to a gentler and non-confrontational tone. The initial four input/output pairs are training examples, and then I tested it with three new inputs:
3
2
1
14
GIF
More examples:
1
0
0
1
GIF
Notice how results are non-deterministic (you can adjust the stochasticity using a parameter), so given the same training examples, and exactly the same prompts, you get different results every time.

1:09 PM · Jul 27, 2020

3
0
0
2
GIF
@gdb Thanks for granting access to the beta! Already exploring different use cases and applications.
0
0
0
0
Replying to @romcabrera
Todo bien?
1
0
0
0
Jajajajja ¿Todo bien en casa amigo?
1
0
0
1
Now in Spanish! It just works out-of-the box, with no additional training, converting sentences from aggresive to a more neutral tone in Spanish. cc: @gdb
0
0
1
3
GIF