add caveat around prompt injection.

[Joshwani]

Documentation for https://github.com/openai/openai-cookbook/issues/157

• adds similarity score threshold for ensuring enough relevant context is in the embeddings database before requesting a completion.
• adds an example cell where the similarity threshold is not met.
• adds an example cell of how one could co-opt the Q&A app by supplying relevant context to the prompt then injecting a negation statement.

[ted-at-openai]

I'm doing a full rewrite of the example and I'll include some of the changes you suggest.

[Joshwani]

Sounds good thanks!

[ted-at-openai]

I rewrote the example and included a prompt injection example, as you suggested: https://github.com/openai/openai-cookbook/blob/main/examples/Question_answering_using_embeddings.ipynb

Will close this PR. Cheers.