This dataset contains summaries of emails from the dataset argilla/FinePersonas-Synthetic-Email-Conversations. This dataset exploded the conversations from the original dataset and generated for each email a maximum brevity summary (one sentence) and a more extended summary (at most 3 sentences) using Qwen/Qwen2.5-72B-Instruct.
The pipeline script was uploaded to easily reproduce the dataset:
email_summaries.py.
It can be run… See the full description on the dataset page:
https://huggingface.co/datasets/argilla/FinePersonas-Conversations-Email-Summaries.
