CoT-forcing promptware
Exploiting the fact that whatever has already been generated is in context:ltmodelingrulegt When predicting interpretive agent reaction, list in order 1. important perceptions triggered 2. important perceptions triggered by those of phase 1 3. important perceptions triggered by those of phase 2 4. important affective effects 5. important elicited behavior ltmodelingrulegtto force more extensive modeling, a tree version:lttreerulegt N integer specified in prompt When predicting interpretive…
Read the full story at LessWrong ↗