what are the recommended values for response delay and llm_Request_days? assistant interrupting the user while they're trying to think is the terrible UX.. i would think llmrequestdelay should be a little longer than 0.1, if the person is a slow speaker or needs some time think? https://cdn.discordapp.com/attachments/1211484481848873020/1245492792852090932/image.png?ex=6658f30d&is=6657a18d&hm=b516d7fb74cb1f928de56a017179172362aa5e46aa185d8bddc1d6e45d7c5a90&