Research map / Large language models / Retrieval and knowledge
RFKG-CoT: Relation-Driven Adaptive Hop-count Selection and Few-Shot Path Guidance for Knowledge-Aware QA
Chao Zhang, Minghan Li, Tianrui Lv, Guodong Zhou
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add relation activation mask to hop-count selector (ablation) | KG-CoT (ChatGPT) · WebQSP | accuracy (%): 82.1 -> 85.5 | Improved |
| add relation activation mask to hop-count selector (ablation) | KG-CoT (ChatGPT) · CompWebQ | accuracy (%): 51.6 -> 59.8 | Improved |
| add few-shot path guidance (ablation) | KG-CoT (ChatGPT) · WebQSP | accuracy (%): 82.1 -> 87.7 | Improved |
| add few-shot path guidance (ablation) | KG-CoT (ChatGPT) · CompWebQ | accuracy (%): 51.6 -> 57.8 | Improved |
| combine mask and few-shot guidance (ablation) | RFKG-CoT (ChatGPT) · WebQSP / CompWebQ | accuracy (%): 87.7 -> 89.9 (WebQSP); 59.8 -> 61.4 (CompWebQ) | Improved |
| replace KG-CoT with RFKG-CoT on small LLM | Llama2-7B · WebQSP | accuracy (%): 72.4 -> 87.1 | Improved |
| replace KG-CoT with RFKG-CoT on GPT-4 | GPT-4 · CompWebQ | accuracy (%): 62.3 -> 65.1 | Improved |
| replace KG-CoT with RFKG-CoT in single-hop setting | ChatGPT · SimpleQuestions | accuracy (%): 77.8 -> 79.2 | Improved |
| increase candidate answer entities K from 5 to 10 (ablation) | RFKG-CoT (ChatGPT) · WebQSP | accuracy (%) | Improved |
| increase K from 10 to 15 (ablation) | RFKG-CoT (ChatGPT) · WebQSP | accuracy (%) | No clear effect |
| increase paths per entity N from 1 to 10 (ablation) | RFKG-CoT (ChatGPT) · WebQSP | accuracy (%) | No clear effect |
| increase few-shot examples beyond 3 (up to 5) (ablation) | RFKG-CoT (ChatGPT) · WebQSP | accuracy (%) | Hurt |
Closest papers by meaning
- Knowledge Editing with Dynamic Knowledge Graphs for Multi-Hop Question Answering (AAAI 2025 · similarity 0.88)
- Debate on Graph: A Flexible and Reliable Reasoning Framework for Large Language Models (AAAI 2025 · similarity 0.88)
- Graph-constrained Reasoning: Faithful Reasoning on Knowledge Graphs with Large Language Models (ICML 2025 · similarity 0.87)
- Reasoning on Graphs: Faithful and Interpretable Large Language Model Reasoning (ICLR 2024 · similarity 0.87)
- Diagnosing and Addressing Pitfalls in KG-RAG Datasets: Toward More Reliable Benchmarking (NeurIPS 2025 · similarity 0.87)
- QFFT, Question-Free Fine-Tuning for Adaptive Reasoning (NeurIPS 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
