Enhancing XAI Narratives through Multi-Narrative Refinement and Knowledge Distillation
Dataset Summary
This collection of datasets has been developed for narrative generation from counterfactuals through knowledge distillation. The main objective is to train small language models (SLMs) to transform structured counterfactual explanations into coherent and human-readable narratives.
The datasets are designed to support a two-stage pipeline:
