Early Evidence of Vibe-Proving with Consumer LLMs: A Case Study on Spectral Region Characterization with ChatGPT-5.2 (Thinking)

2026-02-21

AI Summary

This study investigates the potential of consumer-grade large language models—specifically ChatGPT-5.2 Thinking—as assistants in research-level mathematical proof, with an emphasis on workflows deployable by individual researchers. Focusing on the conjecture by Ran and Teng (2024) concerning the non-real spectral regions of a family of 4-cycle stochastic non-negative matrices, we design an iterative human–AI “generate–review–refine” pipeline that integrates multi-turn dialogue and versioned proof drafts. This approach provides the first empirical demonstration of LLMs’ exploratory utility in high-level “atmospheric” mathematical reasoning. Our work not only establishes necessary and sufficient conditions for the conjecture and constructs corresponding boundary cases, but also highlights the complementary roles of AI in heuristic exploration and human experts in ensuring critical correctness.

Technology Category

• Knowledge Representation and Reasoning: Automated Reasoning and Theorem Proving
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Large Language Models (LLMs) are increasingly used as scientific copilots, but evidence on their role in research-level mathematics remains limited, especially for workflows accessible to individual researchers. We present early evidence for vibe-proving with a consumer subscription LLM through an auditable case study that resolves Conjecture 20 of Ran and Teng (2024) on the exact nonreal spectral region of a 4-cycle row-stochastic nonnegative matrix family. We analyze seven shareable ChatGPT-5.2 (Thinking) threads and four versioned proof drafts, documenting an iterative pipeline of generate, referee, and repair. The model is most useful for high-level proof search, while human experts remain essential for correctness-critical closure. The final theorem provides necessary and sufficient region conditions and explicit boundary attainment constructions. Beyond the mathematical result, we contribute a process-level characterization of where LLM assistance materially helps and where verification bottlenecks persist, with implications for evaluation of AI-assisted research workflows and for designing human-in-the-loop theorem proving systems.

Problem

• Large Language Models
• mathematical proof
• AI-assisted research
• human-in-the-loop
• theorem proving

Innovation

• vibe-proving
• human-in-the-loop theorem proving
• spectral region characterization
• iterative generate-referee-repair pipeline
• consumer LLMs in mathematics