Post #4344486

ooofff...I've got my AI-caution pill for today. I do apply the 'retrieve instead of synthesize'-idea in my daily (local) llm-as-dictionary workflow...i like them as dictionaries and encyclopedias...too bad there isn't a single equivalent textfile of that data (which I could grep instead). Would be cool though: all wikipedia titles + summaries in one textfile. UPDATE: there is! https://www.kaggle.com/datasets/kirauser/wikipedia-dataset