LLM "curving" via prompting

Hey Carter, I don’t think inteligence comes from mathematics but rather is a product of many realtionships and hidden states trained in language, also reasoning. If we constatntly use just token-be-token reasoning we will get nowhere. In the prompt I am providing two differect modes i.e. lenses how the content and with it the hidden language traits (inteligence, reasoning etc.) can potentialy be unlocked. The reasoning is never preditciton based and what self-organizational mode/lend does is inhabit what was predicted with non-linear meaning i.e. strucure. I think if we want to even smell the odour of consciosness or inteligence itself we need to look beyond token-by-token reasoning in linear way and let the system to organize the meaning based on trained relationships. The language is milions of years old and has its own rules that ae governed by real human inteligence. I though think that if we let non-linearity to take over we could research deeper layers of language and inteligence that is embeded in it. Tomaž