KCLarity at SemEval-2026 Task 6: Encoder and Zero-Shot Approaches to Political Evasion Detection

2026-03-06

AI Summary

This study addresses the identification of vagueness and evasiveness in political discourse, aiming to assess statements for clarity or evasiveness. It proposes an “evasiveness-first” hierarchical modeling paradigm: first predicting an evasiveness label and then deriving clarity, enhanced through auxiliary training and zero-shot inference strategies. Experiments employ RoBERTa-large and GPT-5.2, with the latter used for zero-shot reasoning. Results show that RoBERTa-large achieves the best performance on the public test set, while zero-shot GPT-5.2 demonstrates superior generalization on the hidden evaluation set. This work presents the first systematic evaluation of large language models’ zero-shot generalization capabilities for this task, offering a novel paradigm for computational analysis of political discourse.

Technology Category

• Natural Language Processing: Discourse, Pragmatics & Argument Mining
• Machine Learning: Large Multimodal Models (LMMs)
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty

Application Category

• Web Mining and Content Analysis: Models for Web evolution
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

This paper describes the KCLarity team's participation in CLARITY, a shared task at SemEval 2026 on classifying ambiguity and evasion techniques in political discourse. We investigate two modelling formulations: (i) directly predicting the clarity label, and (ii) predicting the evasion label and deriving clarity through the task taxonomy hierarchy. We further explore several auxiliary training variants and evaluate decoder-only models in a zero-shot setting under the evasion-first formulation. Overall, the two formulations yield comparable performance. Among encoder-based models, RoBERTa-large achieves the strongest results on the public test set, while zero-shot GPT-5.2 generalises better on the hidden evaluation set.

Problem

• political evasion
• ambiguity detection
• political discourse
• clarity classification

Innovation

• zero-shot learning
• evasion detection
• hierarchical label inference
• encoder-decoder comparison
• political discourse analysis