List your company
ResearchFormerly WizWand

Knowledge Graph Question Answering on GrailQA (Overall)

86.4Hits@1CoG
25.76841.50957.2572.991Jan 16, 2026
Updated 8mo ago

Evaluation results

MethodLinks
CoGMethod category=Prompting KG-Augmented LLM, Backbone LLM=GPT-42026.0186.4—
KG-AgentMethod category=Fine-Tuned KG-Augmented LLM2026.0186.1—
PoGMethod category=Prompting KG-Augmented LLM, Backbone LLM=GPT-42026.0184.7—
ToGMethod category=Prompting KG-Augmented LLM, Backbone LLM=GPT-42026.0181.4—
CoGMethod category=Prompting KG-Augmented LLM, Backbone LLM=GPT-3.5 or others2026.0179.2—
PoGMethod category=Prompting KG-Augmented LLM, Backbone LLM=GPT-3.5 or others2026.0176.5—
GAINMethod category=Fine-Tuned KG-Augmented LLM2026.0176.3—
PanguMethod category=Fine-Tuned KG-Augmented LLM2026.0175.4—
FC-KBQAMethod category=Fine-Tuned KG-Augmented LLM2026.0173.2—
TIARAMethod category=Fine-Tuned KG-Augmented LLM2026.0173—
CoGMethod category=Prompting KG-Augmented LLM, Backbone LLM=Qwen2.5-7B2026.0172—
PoGMethod category=Prompting KG-Augmented LLM, Backbone LLM=Qwen2.5-7B2026.0168.9—
RnG-KBQAMethod category=Fine-Tuned KG-Augmented LLM2026.0168.8—
ToGMethod category=Prompting KG-Augmented LLM, Backbone LLM=GPT-3.5 or others2026.0168.7—
FlexKBQAMethod category=Fine-Tuned KG-Augmented LLM2026.0162.8—
ToGMethod category=Prompting KG-Augmented LLM, Backbone LLM=Qwen2.5-7B2026.0162.6—
KB-BINDERMethod category=Prompting KG-Augmented LLM, Backbone LLM=GPT-3.5 or others2026.0153.2—
SCMethod category=LLM-Only2026.0129.6—
IO PromptMethod category=LLM-Only2026.0129.4—
CoTMethod category=LLM-Only2026.0128.1—
ChatGPTType=Pure LLMs2025.12—25.3
Davinci-003Type=Pure LLMs2025.12—30.1
GPT-4Type=Pure LLMs2025.12—31.7
KG-AgentType=LLMs + KG2025.12—86.1
PathHDType=LLMs + KG2025.12—86.7
StructGPTType=LLMs + KG2025.12—54.6