List your company
ResearchFormerly WizWand

Multi-Hop Knowledge Graph Question Answering on GrailQA

94.4Hits@1PoG
9.43231.49153.5575.609Oct 18, 2024Jan 1, 2025Mar 17, 2025Jun 1, 2025Aug 15, 2025Oct 29, 2025Jan 13, 2026
Updated 5mo ago

Evaluation results

MethodLinks
PoGClass=ICL, LLM=GPT-4, External Knowledge=With external knowledge2024.1094.4
PoGClass=ICL, LLM=GPT-3.5-Turbo, External Knowledge=With external knowledge2024.1091.6
PoG-EClass=ICL, LLM=GPT-4, External Knowledge=With external knowledge2024.1091.4
PoG-EClass=ICL, LLM=GPT-3.5-Turbo, External Knowledge=With external knowledge2024.1087.6
PrivGemoType=Proposed, Brain LLM=DeepSeek-V3, Hand LLM=Qwen3-32b2026.0186.1
PrivGemoType=Proposed, Brain LLM=GPT-3.5-Turbo, Hand LLM=Qwen3-32b2026.0184.5
PrivGemoType=Proposed, Brain LLM=GPT-4o-mini, Hand LLM=Qwen3-32b2026.0183.6
Logits-to-LogicBackbone LLM=LLaMA3.1-8b2025.1182
ToG/ToG-RClass=ICL, LLM=GPT-4, External Knowledge=With external knowledge2024.1081.4
ToGType=KG-centric RAG, LLM=GPT-42026.0181.4
ToGMethod Paradigm=Agentic Reasoning, Backbone LLM=GPT42025.1181.4
ToG-RMethod Paradigm=Agentic Reasoning, Backbone LLM=GPT42025.1180.3
DoGMethod Paradigm=Agentic Reasoning2025.1180
ARoGType=Privacy-aware KGQA, LLM=GPT-4o-mini2026.0178.7
DARAMethod Paradigm=Agentic Reasoning, Backbone LLM=LLaMA2-13b2025.1177
PoGType=KG-centric RAG, LLM=GPT-3.5-Turbo2026.0176.5
PoGMethod Paradigm=Agentic Reasoning, Backbone LLM=ChatGPT2025.1176.5
Prior FT SOTAClass=SL, External Knowledge=With external knowledge2024.1075.4
ToG/ToG-RClass=ICL, LLM=GPT-3.5-Turbo, External Knowledge=With external knowledge2024.1068.7
ToGType=KG-centric RAG, LLM=GPT-3.5-Turbo2026.0168.7
ToGMethod Paradigm=Agentic Reasoning, Backbone LLM=ChatGPT2025.1168.7
KB-BINDERClass=ICL, LLM=Codex, External Knowledge=With external knowledge2024.1058.5
ToG-RMethod Paradigm=Agentic Reasoning, Backbone LLM=ChatGPT2025.1156.4
o3-mini2025.0543.8
Qwen2.5-72B2025.0536.1
QwQ-32B2025.0535.8
Qwen2.5-32B + fs1Protocol=fs12025.0534.4
Qwen2.5-32B + rtProtocol=rt2025.0534.2
R1-70B2025.0534
Qwen2.5-32B2025.0533.4
Qwen2.5-32B + cotProtocol=CoT2025.0533.2
Qwen2.5-14B + rtProtocol=rt2025.0533.1
Qwen2.5-14B + fs1Protocol=fs12025.0532.8
Qwen2.5-14B + cotProtocol=CoT2025.0532.2
Qwen2.5-14B2025.0531.8
Qwen2.5-7B + fs1Protocol=fs12025.0530.3
Qwen2.5-7B + rtProtocol=rt2025.0530
Qwen2.5-7B2025.0529.9
SCLLM=GPT-3.5-Turbo, External Knowledge=Without external knowledge2024.1029.6
SCType=LLM-only, LLM=GPT-3.5-Turbo2026.0129.6
SCMethod Paradigm=LLMs Reasoning, Backbone LLM=ChatGPT2025.1129.6
Qwen2.5-7B + cotProtocol=CoT2025.0529.5
IO promptLLM=GPT-3.5-Turbo, External Knowledge=Without external knowledge2024.1029.4
IO promptType=LLM-only, LLM=GPT-3.5-Turbo2026.0129.4
IO promptMethod Paradigm=LLMs Reasoning, Backbone LLM=ChatGPT2025.1129.4
CoTLLM=GPT-3.5-Turbo, External Knowledge=Without external knowledge2024.1028.1
CoTType=LLM-only, LLM=GPT-3.5-Turbo2026.0128.1
CoTMethod Paradigm=LLMs Reasoning, Backbone LLM=ChatGPT2025.1128.1
Qwen2.5-3B + rtProtocol=rt2025.0527.9
Qwen2.5-3B + fs1Protocol=fs12025.0525.3
Qwen2.5-3B2025.0525.2
Qwen2.5-3B + cotProtocol=CoT2025.0524.8
SmolLM2-1.7B + rtProtocol=rt2025.0522.3
SmolLM2-1.7B2025.0521.9
SmolLM2-1.7B + fs1Protocol=fs12025.0521.8
Qwen2.5-1.5B + cotProtocol=CoT2025.0521.6
Qwen2.5-1.5B + rtProtocol=rt2025.0521.2
SmolLM2-1.7B + cotProtocol=CoT2025.0520.9
Qwen2.5-1.5B2025.0520.8
Qwen2.5-1.5B + fs1Protocol=fs12025.0520.4
SmolLM2-360M + cotProtocol=CoT2025.0516.9
Qwen2.5-0.5B + fs1Protocol=fs12025.0516.6
SmolLM2-360M2025.0516.4
SmolLM2-360M + rtProtocol=rt2025.0515.6
Qwen2.5-0.5B + rtProtocol=rt2025.0515.5
Qwen2.5-0.5B + cotProtocol=CoT2025.0514.1
SmolLM2-360M + fs1Protocol=fs12025.0513.8
Qwen2.5-0.5B2025.0512.7