{"ID":2879412,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2508.16705","arxiv_id":"2508.16705","title":"Assessing Consciousness-Related Behaviors in Large Language Models Using the Maze Test","abstract":"We investigate consciousness-like behaviors in Large Language Models (LLMs) using the Maze Test, challenging models to navigate mazes from a first-person perspective. This test simultaneously probes spatial awareness, perspective-taking, goal-directed behavior, and temporal sequencing-key consciousness-associated characteristics. After synthesizing consciousness theories into 13 essential characteristics, we evaluated 12 leading LLMs across zero-shot, one-shot, and few-shot learning scenarios. Results showed reasoning-capable LLMs consistently outperforming standard versions, with Gemini 2.0 Pro achieving 52.9% Complete Path Accuracy and DeepSeek-R1 reaching 80.5% Partial Path Accuracy. The gap between these metrics indicates LLMs struggle to maintain coherent self-models throughout solutions -- a fundamental consciousness aspect. While LLMs show progress in consciousness-related behaviors through reasoning mechanisms, they lack the integrated, persistent self-awareness characteristic of consciousness.","short_abstract":"We investigate consciousness-like behaviors in Large Language Models (LLMs) using the Maze Test, challenging models to navigate mazes from a first-person perspective. This test simultaneously probes spatial awareness, perspective-taking, goal-directed behavior, and temporal sequencing-key consciousness-associated chara...","url_abs":"https://arxiv.org/abs/2508.16705","url_pdf":"https://arxiv.org/pdf/2508.16705v1","authors":"[\"Rui A. Pimenta\",\"Tim Schlippe\",\"Kristina Schaaff\"]","published":"2025-08-22T11:14:36Z","proceeding":"cs.CL","tasks":"[\"cs.CL\",\"cs.AI\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false}