Custom Moderation GPT Model | Fine Tuning

[jr.2509]

While I fully support @dignity_for_all’s suggestion, I also recalled a similar question from a user some time ago. At the time, I shared a few ideas for the training data composition. I did not hear back from the user in question, so can’t confirm if the approach ended up being successful. For what it’s worth, I am sharing the link to the post anyway:

Avoid certain responses and prompts and generate responses as per my input

Thanks for sharing this example. Overall, this example is suitable for training purposes. Note that you can remove the phrase “DIRECTIVE_FOR_gpt-3.5-turbo-1106 :” - this is already implied as it is a system message. As I don’t have line of sight into your broader training data set, here are a couple of general points. This takes into account some of the earlier messages in this thread: Make sure to use a consistent system message in your training data. In this example you provided it seems …