Co-Sight: Enhancing LLM-Based Agents via Conflict-Aware Meta-Verification and Trustworthy Reasoning with Structured Facts

2025-10-24

AI Summary

LLMs often fail in long-horizon reasoning not due to insufficient generative capacity, but because of inadequate verification of intermediate reasoning steps. This paper proposes a conflict-aware meta-verification framework that targets reasoning divergence points for efficient, goal-directed falsification. It integrates structured fact management with cross-agent evidence synchronization to establish an auditable, falsifiable, closed-loop collaborative reasoning paradigm. Key contributions are: (1) conflict-driven verification, which substantially reduces redundant computation; and (2) modularization of knowledge into verifiable factual units, enhancing transparency and traceability of reasoning. The framework achieves state-of-the-art accuracy—84.4% on GAIA, 35.5% on Humanity’s Last Exam, and 93.8% on Chinese SimpleQA—outperforming all existing methods across benchmarks.

Technology Category

• Knowledge Representation and Reasoning: Reasoning with Beliefs
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Long-horizon reasoning in LLM-based agents often fails not from generative weakness but from insufficient verification of intermediate reasoning. Co-Sight addresses this challenge by turning reasoning into a falsifiable and auditable process through two complementary mechanisms: Conflict-Aware Meta-Verification (CAMV) and Trustworthy Reasoning with Structured Facts (TRSF). CAMV reformulates verification as conflict identification and targeted falsification, allocating computation only to disagreement hotspots among expert agents rather than to full reasoning chains. This bounds verification cost to the number of inconsistencies and improves efficiency and reliability. TRSF continuously organizes, validates, and synchronizes evidence across agents through a structured facts module. By maintaining verified, traceable, and auditable knowledge, it ensures that all reasoning is grounded in consistent, source-verified information and supports transparent verification throughout the reasoning process. Together, TRSF and CAMV form a closed verification loop, where TRSF supplies structured facts and CAMV selectively falsifies or reinforces them, yielding transparent and trustworthy reasoning. Empirically, Co-Sight achieves state-of-the-art accuracy on GAIA (84.4%) and Humanity's Last Exam (35.5%), and strong results on Chinese-SimpleQA (93.8%). Ablation studies confirm that the synergy between structured factual grounding and conflict-aware verification drives these improvements. Co-Sight thus offers a scalable paradigm for reliable long-horizon reasoning in LLM-based agents. Code is available at https://github.com/ZTE-AICloud/Co-Sight/tree/cosight2.0_benchmarks.

Problem

• Addresses unreliable long-horizon reasoning in LLM-based agents
• Enhances verification through conflict identification and structured facts
• Improves reasoning accuracy and transparency via falsifiable processes

Innovation

• Conflict-Aware Meta-Verification targets reasoning disagreement hotspots
• Trustworthy Reasoning organizes evidence via structured facts module
• Closed verification loop combines structured facts with selective falsification