List your company
ResearchFormerly WizWand

Question Answering on HotpotQA (LLM-Acc and Cont-Acc)

94.5LLM AccuracyA-RAG (Full)
30.74847.29963.8580.401Feb 3, 2026
Updated 8mo ago

Evaluation results

MethodLinks
A-RAG (Full)Backbone LLM=GPT-5-mini2026.0294.588
A-RAG (Naive)Backbone LLM=GPT-5-mini2026.0290.885.3
LinearRAGBackbone LLM=GPT-5-mini2026.0286.277.6
HippoRAG2Backbone LLM=GPT-5-mini2026.0284.875
GraphRAGBackbone LLM=GPT-5-mini2026.0282.574.9
Naive RAGBackbone LLM=GPT-5-mini2026.0281.279.5
HippoRAG2Backbone LLM=GPT-4o-mini2026.0280.769.7
A-RAG (Full)Backbone LLM=GPT-4o-mini2026.0277.174
FaithfulRAGBackbone LLM=GPT-5-mini2026.0276.975.3
A-RAG (Naive)Backbone LLM=GPT-4o-mini2026.0276.670.7
Naive RAGBackbone LLM=GPT-4o-mini2026.0274.572.9
LinearRAGBackbone LLM=GPT-4o-mini2026.027260.5
MA-RAGBackbone LLM=GPT-5-mini2026.0267.157.9
Direct AnswerBackbone LLM=GPT-5-mini2026.0263.653.5
RAGentABackbone LLM=GPT-4o-mini2026.026362.4
RAGentABackbone LLM=GPT-5-mini2026.0261.265
MA-RAGBackbone LLM=GPT-4o-mini2026.0260.654.4
FaithfulRAGBackbone LLM=GPT-4o-mini2026.0260.552.5
Direct AnswerBackbone LLM=GPT-4o-mini2026.0245.440.7
GraphRAGBackbone LLM=GPT-4o-mini2026.0233.233.3