. . . . "KG-o1: Enhancing Multi-hop Question Answering in Large Language Models via Knowledge Graph Integration" . . . . . . . . . . . . "ChatGPT-4o is an advanced general-purpose large language model (GPLLM) used in two contexts: first, as a tool to generate multi-hop questions for the KG-MHQA SFT dataset creation, and second, as a powerful baseline for performance comparison in the experiments." . "ChatGPT-4o" . . "ChatGPT4o-mini is an advanced general-purpose large language model (GPLLM) used as a baseline model for comparative evaluation against the proposed KG-o1 models on multi-hop reasoning tasks." . "ChatGPT4o-mini" . . "DeepSeek-R1 is a specific large reasoning model (LRM) used as a strong baseline for evaluating the performance of KG-o1 models on multi-hop question answering datasets." . "DeepSeek-R1" . . "GRPO is a reinforcement learning method that serves as a comparative baseline in the ablation studies, where its performance in boosting LLMs' multi-hop reasoning is contrasted with other fine-tuning and optimization strategies, including the paper's Self-improved Adaptive DPO." . "GRPO" . . "Gemini 2.0 Flash Thinking is a specific large reasoning model (LRM) used as a strong baseline for evaluating the performance of KG-o1 models on multi-hop question answering tasks, highlighting its advanced reasoning capabilities." . "Gemini 2.0 Flash Thinking" . . . "KG-o1 is a novel four-stage framework that integrates Knowledge Graphs (KGs) to enhance Large Language Models' (LLMs) multi-hop reasoning abilities. It involves constructing KG-derived datasets (KG-MHQA SFT and DPO) and using them to fine-tune LLMs (via Supervised Fine-Tuning and a \"Self-improved Adaptive DPO\" strategy), aiming to improve the LLM's intrinsic knowledge expression and reasoning capabilities during a training stage by internalizing logical paths." . "KG-o1" . . . "o1-mini is a specific large reasoning model (LRM) mentioned as a prominent baseline for comparison against the proposed KG-o1 models in multi-hop question answering tasks." . "o1-mini" . . "Open-o1 is a specific influential large reasoning model (LRM) from the open-source community, used as a baseline for performance comparison in the experiments of the paper." . "Open-o1" . . "QwQ-32B-Preview is a specific large reasoning model (LRM) included as a prominent baseline for comparative experiments against the KG-o1 models on multi-hop reasoning tasks." . "QwQ-32B-Preview" . . . "2026-03-13T16:10:14.372Z"^^ . . . . "LLM-KG assessment for paper 10.48550/arXiv.2508.15790" . "RSA" . "MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAwNz2QK3SEifno78S7+48zUB0xpTex3mAzW73ZimHqNcdEMU5/apslrGrTHGFAt/Chocgo++r6JQp5ygY7NyJHGWdaIqnt85pjX4PbNfLAvapyUO00qZP34fY61w4eZ9UMtleWEsmZKRtQPyJ8ODl46i/rfPuZlcJGpM9Nmy5mpGWuepqIEvF4a/t7pLVeCEDFSYXT+yaiygt6ynIK5f7TtEDhZpeUf/Q74WhMPJXm4yTU/hqOX4IW+50kWHNArGGZwUaXwzyG6M3Zd6UMModryGkLqS4H/MSE3ZA1Ylnms7BfWLEXhMWlaKi6HRV4nGRDLhxVSi9LSRi3LWKLhNIIQIDAQAB" . "JwesqkdoglWm3/mXA9J8rFpVgW/WxZ4BNq6HnvpmaPwF5MwyNA/dauZGFS67c5xMKrrXdr/AbqsEfPIZ6ZazcmWhu0AVKP/xnFHPbKJXFlS+Pl2jufAYGJ50ZSxXX88mv3L5H9IovO/SjdUo9Wa/VRIc/PlVFgn133b10hOM8YitFc3hRufK/FSce7FdKD4KJauXmPX/g7tTa+Usxcdmse7pm3zeGu8PjS8Ojn+FH6gexIo8BMeZDtWeYlOyAo/GZSq4+IrAbE1JUVz9xU5rVaf7NzCMNwLUvX0jgt8L1yD8E/3N/nvQ7jpLKkt0bWbNM23jQF2XWkiMLGKMFLxnRQ==" . . .