[2603.19849] Semantic Delta: An Interpretable Signal Differentiating Human and LLMs Dialogue

Computer Science > Computation and Language arXiv:2603.19849 (cs) [Submitted on 20 Mar 2026] Title:Semantic Delta: An Interpretable Signal Differentiating Human and LLMs Dialogue Authors:Riccardo Scantamburlo, Mauro Mezzanzana, Giacomo Buonanno, Francesco Bertolotti View a PDF of the paper titled Semantic Delta: An Interpretable Signal Differentiating Human and LLMs Dialogue, by Riccardo Scantamburlo and 3 other authors View PDF HTML (experimental) Abstract:Do LLMs talk like us? This question intrigues a multitude of scholar and it is relevant in many fields, from education to academia. This work presents an interpretable statistical feature for distinguishing human written and LLMs generated dialogue. We introduce a lightweight metric derived from semantic categories distribution. Using the Empath lexical analysis framework, each text is mapped to a set of thematic intensity scores. We define semantic delta as the difference between the two most dominant category intensities within a dialogue, hypothesizing that LLM outputs exhibit stronger thematic concentration than human discourse. To evaluate this hypothesis, conversational data were generated from multiple LLM configurations and compared against heterogeneous human corpora, including scripted dialogue, literary works, and online discussions. A Welch t-test was applied to the resulting distributions of semantic delta values. Results show that AI-generated texts consistently produce higher deltas than human texts, indi...