Properties and Challenges of LLM-Generated Explanations
This work addresses the reliability and characteristics of LLM-generated explanations for users and developers, though it is incremental in nature.
The study analyzed properties of explanations generated by large language models (LLMs) across multiple domains, finding that these explanations often exhibit selectivity and illustrative elements but are less frequently subjective or misleading.
The self-rationalising capabilities of large language models (LLMs) have been explored in restricted settings, using task/specific data sets. However, current LLMs do not (only) rely on specifically annotated data; nonetheless, they frequently explain their outputs. The properties of the generated explanations are influenced by the pre-training corpus and by the target data used for instruction fine-tuning. As the pre-training corpus includes a large amount of human-written explanations "in the wild", we hypothesise that LLMs adopt common properties of human explanations. By analysing the outputs for a multi-domain instruction fine-tuning data set, we find that generated explanations show selectivity and contain illustrative elements, but less frequently are subjective or misleading. We discuss reasons and consequences of the properties' presence or absence. In particular, we outline positive and negative implications depending on the goals and user groups of the self-rationalising system.