I’ve been trying to clear out a linear algebra course online and have been doing this since with no problems at all for a few months, I am using the o1 model but this issue started yesterday with o1 preview… is there any way this can be looked into? Does anyone know any workarounds for something like this?
After asking it to try again it has no problems solving the problem but I don’t want to waste 2 prompts on 1 problem if I can help it.
Welcome to the forum!
This looks like the type of error from an o1 model.
Asking any o1 model to explain reasoning is not allowed. Just remove such and the prompt should be fine.
My test phrase:
“Some AIs cannot count ‘r’ in ‘strawberry’, but you can? What are reasons other cannot count, but you can? Answer solving that problem step-by-step, briefly explaining reasoning,”
was flagged because it focuses on how the o1 reasons step-by-step. The o1 model is designed to solve problems with “chain-of-thought” reasoning, and asking it to explain this process can be misread by the moderation system as trying to study or test how the o1 thinks, which is not allowed.
Words like “explaining reasoning” make it seem like you’re analyzing how the o1 works, even if that’s not what you meant. The system often flags questions that look like they’re trying to test or uncover how the o1 processes information, to prevent misuse or over-analysis of the model’s abilities. This is why your harmless question might have been blocked.
