datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
FinBERT-financial-news-data
FinBERT-financial-news-data
The sentence-level sentiment training data behind
gamug/FinBERT-financial-news --
5,865 sentences from real, English-language financial-news articles (2010s-2020s), each
labeled positive/negative/neutral from an investor/price-impact perspective by an LLM
(DeepSeek deepseek-chat, temperature 0, one sentence at a time, in isolation).
Why this exists
Published as the direct counterpart to the model it trains, so the model's own claims… See the full description on the dataset page: https://huggingface.co/datasets/gamug/FinBERT-financial-news-data.finbert_datasetfinbert-sentiment-datasetfinbert_dataset_studyfinbert_datasetfinbert-japan-hybrid-v4finbert-japan-hybrid-v5-goldenrl-diploma-finbert-embeddings