Content filter overreaction is making work difficult | Voters

Content filter overreaction is making work difficult

complete

EvaL

The content filter is driving me crazy. It gets triggered by the most harmless phrases, making it really hard to get anything done.

I can make the exact same request directly in ChatGPT, and everything flows smoothly. This is a Merlin issue, not a problem with the models themselves.

March 16, 2025

Siddhartha

marked this post as

complete

We have reduced the scope of filters.

Renji人機

Siddhartha Hi, it seems that the filters are still very strong for fictional writings. I can't produce any of my work for novels as it is still being flagged even if it a normal romantic novel.

I've tried Chatgpt and POE using the same prompt and they are able to generate the outcome without much of an issue.

Siddhartha

Renji人機: It might be flagged as sexual content. We are going to remove those from the filter for the website in some time. After that only sexual content related to children will be banned.

Renji人機

Siddhartha Thank you, if possible please tone down the violence filters as well, or give us an option to bypass the filters so that it could smoothen our fictional work. Thank you.