Probing Commonsense Explanation in Dialogue Response Generation

TL;DR

This paper investigates how well dialogue response generation models understand commonsense reasoning by analyzing their ability to produce and interpret explanations, revealing significant gaps in their logical understanding.

Contribution

It formalizes commonsense as a latent variable in response generation and introduces annotated explanations and probing methods to evaluate models' CSR capabilities.

Findings

Models struggle to capture logical relations in explanations

Fine-tuning and larger models do not improve CSR understanding

Proposed probing settings reveal gaps in models' reasoning abilities

Abstract

Humans use commonsense reasoning (CSR) implicitly to produce natural and coherent responses in conversations. Aiming to close the gap between current response generation (RG) models and human communication abilities, we want to understand why RG models respond as they do by probing RG model's understanding of commonsense reasoning that elicits proper responses. We formalize the problem by framing commonsense as a latent variable in the RG task and using explanations for responses as textual form of commonsense. We collect 6k annotated explanations justifying responses from four dialogue datasets and ask humans to verify them and propose two probing settings to evaluate RG models' CSR capabilities. Probing results show that models fail to capture the logical relations between commonsense explanations and responses and fine-tuning on in-domain data and increasing model sizes do not lead to understanding of CSR for RG. We hope our study motivates more research in making RG models emulate the human reasoning process in pursuit of smooth human-AI communication.