Beyond the Black Box: Demystifying Multi-Turn LLM Reasoning with VISTA
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (48) · 41745-41747
DOI: 10.1609/aaai.v40i48.42401 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Yiran Zhang und 3 weiteren Autor:innen.
Abstract (Originalsprache)
Recent research has increasingly focused on the reasoning capabilities of Large Language Models (LLMs) in multi-turn interactions, as these scenarios more closely mirror real-world problem-solving. However, analyzing the intricate reasoning processes within these interactions presents a significant challenge due to complex contextual dependencies and a lack of specialized visualization tools, leading to a high cognitive load for researchers. To address this gap, we present VISTA, an web-based Visual Interactive System for Textual Analytics in multi-turn reasoning tasks. VISTA allows users to visualize the influence of context on model decisions and interactively modify conversation histories to conduct "what-if" analyses across different models. Furthermore, the platform can automatically parse a session and generate a reasoning dependency tree, offering a transparent view of the model's step-by-step logical path. By providing a unified and interactive framework, VISTA significantly reduces the complexity of analyzing reasoning chains, thereby facilitating a deeper understanding of the capabilities and limitations of current LLMs. The platform is open-source and supports easy integration of custom benchmarks and local models.
Zitieren
APA
Zhang, Y., Lin, M., Dras, M., & Naseem, U. (2026). Beyond the Black Box: Demystifying Multi-Turn LLM Reasoning with VISTA. Proceedings of the AAAI Conference on Artificial Intelligence, 40(48), 41745-41747. https://doi.org/10.1609/aaai.v40i48.42401
Vancouver
Zhang Y, Lin M, Dras M, Naseem U. Beyond the Black Box: Demystifying Multi-Turn LLM Reasoning with VISTA. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(48):41745-41747. doi:10.1609/aaai.v40i48.42401
BibTeX
@article{Zhang2026,
author = {Zhang, Yiran and Lin, Mingyang and Dras, Mark and Naseem, Usman},
title = {Beyond the Black Box: Demystifying Multi-Turn LLM Reasoning with VISTA},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {48},
pages = {41745--41747},
doi = {10.1609/aaai.v40i48.42401}
}
EUVIMED