every time i send a request to this model, it just cuts off randomly. in the middle of sentences, words, whatever. sometime it can’t even get one word out before stopping. whats up with this?
Perhaps try changing safety settings to ‘block none’. Those filters can be over sensitive.
@Andy_Sneddon Thank you for the feedback.
We tried with Gemini flash 2.0 with some random request , it is working fine.
Still, to understand more and reproduce the issue, could you please provide some more details on the request?
Thank you.
this specific instance isnt happening anymore, but i feel like safety filters still apply to the thinking. like if i were to disable thinking it would output something, but if i enable thinking it will say something and get blocked that wouldnt get blocked otherwise.