Humans often refer to personal narratives, life experiences, and events to\nmake a conversation more engaging and rich. While persona-grounded dialog\nmodels are able to generate responses that follow a given persona, they often\nmiss out on stating detailed experiences or events related to a persona, often\nleaving conversations shallow and dull. In this work, we equip dialog models\nwith 'background stories' related to a persona by leveraging fictional\nnarratives from existing story datasets (e.g. ROCStories). Since current dialog\ndatasets do not contain such narratives as responses, we perform an\nunsupervised adaptation of a retrieved story for generating a dialog response\nusing a gradient-based rewriting technique. Our proposed method encourages the\ngenerated response to be fluent (i.e., highly likely) with the dialog history,\nminimally different from the retrieved story to preserve event ordering and\nconsistent with the original persona. We demonstrate that our method can\ngenerate responses that are more diverse, and are rated more engaging and\nhuman-like by human evaluators, compared to outputs from existing dialog\nmodels.\n