Submissions are evaluated based on the quality of the generated text compared to the original content. We primarily use the BLEU score and Jaccard similarity to measure the coherence, relevance, and fluency of the generated content.

## Submission File

For each ID in the test set, you must generate a corresponding text content. The file should contain a header and have the following format:


Id,label
1,"Generated content for ID 1..."
2,"Generated content for ID 2..."
3,"Generated content for ID 3..."
etc.


Please ensure that the generated content is relevant and closely aligned with the context of the given dataset.