并行大语言模型推理用于偏见鲁棒、稳健的概念抽象
Parallel LLM Reasoning for Bias-Resilient, Robust Conceptual Abstraction
AI总结:
本文提出了一种结合并行分块处理与证据锚定整合的结构化框架,旨在减少长文档分析中的偏见、遗漏误差和过度泛化问题,通过并行处理和证据锚定提高文本分析的可靠性和可扩展性。
摘要:
大型语言模型(LLMs)在分析文本方面被越来越多地使用。然而,当分析长文档时,它们常常受到上下文推理限制的困扰。当长文档被顺序处理时,早期或主导的概念会掩盖不明显但有意义的解释,导致累积分析偏见、遗漏误差和过度泛化。此外,独立生成的输出通常在没有系统基础的情况下合并,引入了冗余、概念漂移和未经支持的主张。本研究提出了一种结合并行分块处理与证据锚定整合的结构化框架。文本首先被划分为语义连贯的分块,并独立并行处理以消除早期处理的影响。然后,独立生成的解释通过显式的证据锚定和优先级整合进行整合,从而减少主导和过度泛化,同时提高可追溯性。在多种模型类型和规模上的实验表明,并行处理显著减少了约84%的遗漏误差,提高了高达130%的证据可追溯性,并减少了高达91%的未经支持的主张。较小的模型受益最大,表明高效的并行分块和整合在实现可靠和可扩展的文本分析中起关键作用。
英文摘要:
Large language models (LLMs) have been increasingly used to analyze text. However, they are often plagued with contextual reasoning limitations when analyzing long documents. When long documents are processed sequentially, early or dominant concepts can overshadow less visible but meaningful interpretations, leading to cumulative analytical bias, omission error, and over-generalization. Additionally, independently generated outputs are often merged without systematic grounding, introducing redundancy, conceptual drift, and unsupported claims. This study proposes a structured framework combining parallel chunk-level processing with evidence-anchored consolidation. Texts are first divided into semantically coherent chunks and processed independently in parallel to remove influence from earlier processing. The independently generated interpretations are then consolidated using explicit evidence anchoring and prioritization that reduces dominance and over-generalization while improving traceability. Experiments with multiple model types and sizes indicate that parallel processing significantly reduces omission error by approximately 84%, increases evidence traceability by up to 130%, and reduces unsupported claims by up to 91%. Smaller models benefited most, suggesting that efficient parallel chunking and consolidation play a critical role in achieving reliable and scalable textual analysis.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。