Feeding the AI model a larger text only highlights a few of the correct sentences. This can be improved by tweaking the training data, input data, context-to-sentence-ratio, or other parameters.
Since we are moving to embeddings this can be fixed later. It will also be possible to simply divide the data into paragraphs if longer texts need analyzing. Added label #Wontfix until further development.
Feeding the AI model a larger text only highlights a few of the correct sentences. This can be improved by tweaking the training data, input data, context-to-sentence-ratio, or other parameters.