Teaching ChatGPT mental models/heuristics from that philosophy paper.
- For phil heuristics, rephrase a claim 5 times and see which are really equivalent, mess up the syntax as much as you can. Especially useful where you seem to have hit a dead end and need to wring out an extra distinction.
Ask perplexity which prompts or scaffolds people use to make LLMs produce more novel ideas in experiments
Just asking, what are some asymmetries (e.g. praise/blame) here may be a useful prompt. Feeding in Alexander Pruss’s blog posts might be a good option
See also Mental models