Book-excerpt summarization at seven context lengths (2K – 256K tokens), built for long-context supervised fine-tuning and context-length stress-testing. Each example asks a model to summarize a passage within a target word count; the reference summary was generated by an LLM (see Provenance).

from datasets import load_dataset

ds =… See the full description on the dataset page: https://huggingface.co/datasets/lefft/excerpt-summary-longctx.