Chain of Density
Chain of density is a summarization technique in which the model iteratively rewrites a summary, each pass adding more salient entities while keeping total length constant. The first draft is usually sparse and entity-light; each subsequent pass identifies entities missing from the previous version and folds them in by compressing or rephrasing existing sentences. The result is a dense, entity-rich summary that fits a fixed length budget. It was introduced for news summarization and works best when the target summary must cover all salient entities within a strict word count. It is less useful when the goal is a loose overview rather than an information-dense abstract.
Example
A legal team asks a model to produce a 120-word summary of a 40-page contract. The first pass covers three parties and the effective date. The second pass compresses boilerplate phrasing to add the indemnity cap and governing law. The third pass trims adjectives to fit the termination clause and auto-renewal window. Final length stays at 120 words but entity coverage roughly doubles across the three passes.
Put this into practice
Build polished, copy-ready prompts in under 60 seconds with SurePrompts.
Try SurePrompts