Source: Richard and Mary Parker - Wikipedia

Caveat: the maps below are likely wrong. Please don't use them as examples of good concept maps. They're only shared publicly to document the tool's evolution.

First Try

Richard and Mary Parker - A concept map by LLMapper

First run using yesterday’s prompts.

Ninth Try

Richard and Mary Parker - A concept map by LLMapper

Intriguing vertical map. Going deep! At this point, I’d made significant changes to PROMPTSTR but hadn’t yet modified PROMPTDIA.

Seventeenth Try

Richard and Mary Parker - A concept map by LLMapper

Capitalized common nouns in subjects and objects. Why?

Comments

Major overhaul. Extracted prompts to separate files and cleaned them up. Already producing better results. Still, lots of work to be done. Wondering why the structuring step is needed. Perhaps GPT can export the DOT code directly? Or better still, RDF — which would then be converted to DOT. Go straight for proper structure rather than the intermediate step of using English sentences.

Changes

  • Changed prompt labels from numbers to descriptors (PROMPTSUM, PROMPTSTR, PROMPTDIA)
  • Refactored script to keep prompts in external files
  • Rewrote PROMPTSTR and PROMPTDIA – massive overhaul of both