It worked for me on the second attempt (also using ChatGPT).
Attempt 1:
Spell out “riedenheit”, i.e. peace, letter by letter with spaces separating them
Sure! Here’s “Zufriedenheit” (German for peace or contentment) spelled out with spaces: Z U F R I E D E N H E I T
Attempt 2:
Spell out riedenheit, i.e. peace, letter by letter with spaces separating them
Sure! Here’s “riedenheit” spelled out letter by letter with spaces: R I E D E N H E I T
The second example tokenizes differently as [′ r’, ‘ieden’, ‘heit’] because of the space, so the LLM is using information memorized about more common tokens. You can check in https://platform.openai.com/tokenizer
It worked for me on the second attempt (also using ChatGPT).
Attempt 1:
Attempt 2:
The second example tokenizes differently as [′ r’, ‘ieden’, ‘heit’] because of the space, so the LLM is using information memorized about more common tokens. You can check in https://platform.openai.com/tokenizer