HuggingFaceFW/finephrase

Dataset Card for HuggingFaceFW/finephrase Dataset Summary Synthetic data generated by DataTrove: Model: HuggingFaceTB/SmolLM2-1.7B-Instruct (main) Source dataset: HuggingFaceFW/fineweb-edu, config sample-350BT, split train Generation config: temperature=1.0, top_p=1.0, top_k=50, max_tokens=2048, model_max_context=8192 Speculative decoding: {"method":"suffix","num_speculative_tokens":32} System prompt: None Input column: text Prompt families: faq prompt Rewrite… See the full description on the dataset page: https://huggingface.co/datasets/HuggingFaceFW/finephrase.