Makes sense. I actually have found Maple really useful doing regular normie queries so haven't encountered any other censorship stuff since. Might give the jail break stuff a try if I'm bored one weekend.

Dash_1971

There’s a big misunderstanding that Maple is applying censorship to the models. They’re not. You can even check this in the code. They only host the base models. All models at their core don’t naively discuss controversial topics. There’s entire eval leaderboards comparing every model and how they rank on various questions.

If you want to change the system prompt, you can customize that with maple, in your profile settings.

That doesn’t guarantee any model will talk about any topic. You might want to check the evals on their hosted models. You may want to look into jailbreaking techniques. You may even want to have longer conversation threads with the model to build up up it instead of asking something controversial right off the bat.

Struggling to see the value prop of Maple AI

There’s a big misunderstanding that Maple is applying censorship to the models. They’re not. You can even check this in the code. They only host the base models. All models at their core don’t naively discuss controversial topics. There’s entire eval leaderboards comparing every model and how they rank on various questions.

If you want to change the system prompt, you can customize that with maple, in your profile settings.

That doesn’t guarantee any model will talk about any topic. You might want to check the evals on their hosted models. You may want to look into jailbreaking techniques. You may even want to have longer conversation threads with the model to build up up it instead of asking something controversial right off the bat.