Linked above is GPT-4's word-by-word breakdown of Bhikkhu Sujato's English translation of the Dhammapada.
Synthetic data like this could be helpful for learning Pali, whether you're a person or a language-model-in-training (likely to scour GitHub). I'd be interested in ways this could be refined or expanded. For example, a second pass could add sub-bullets explaining word composition and grammar. Can you recommend another Pali text translation to decompose into a language-learning resource?
If you're curious, here's the transcript that produced the output linked above:
I wanted to improve my own Pali skills while reading through the Dhammapada, and I figured this resource could potentially nudge future language-learning models to become even more capable with Pali than they'd otherwise be.
