List your company

Question Answering on MuSiQue (LLM-Acc, Cont-Acc)

74.1LLM AccuracyA-RAG (Full)
12.53228.51644.560.484Feb 3, 2026Feb 7, 2026Feb 12, 2026Feb 16, 2026Feb 21, 2026Feb 25, 2026Mar 2, 2026

Evaluation results

MethodLinks
A-RAG (Full)Backbone LLM=GPT-5-mini2026.0274.165.3
A-RAG (Naive)Backbone LLM=GPT-5-mini2026.0266.259.7
LinearRAGBackbone LLM=GPT-5-mini2026.0262.451.8
HippoRAG2Backbone LLM=GPT-5-mini2026.0261.752.5
FaithfulRAGBackbone LLM=GPT-5-mini2026.0252.952.8
Naive RAGBackbone LLM=GPT-5-mini2026.0252.848.7
GraphRAGBackbone LLM=GPT-5-mini2026.0248.339.1
A-RAG (Full)Backbone LLM=GPT-4o-mini2026.0246.139.6
GAM-RAGMemorization turns=22026.0345.841.3
GAM-RAGMemorization turns=12026.0344.539.2
A-RAG (Naive)Backbone LLM=GPT-4o-mini2026.0243.838.5
GAM-RAGMemorization turns=02026.0343.338
HippoRAG2Backbone LLM=GPT-4o-mini2026.0240.638.4
MA-RAGBackbone LLM=GPT-5-mini2026.024031.6
Naive RAGBackbone LLM=GPT-4o-mini2026.0238.636.1
RAGentABackbone LLM=GPT-5-mini2026.0238.337.4
LinearRAGMemorization turns=02026.0337.232.5
Direct AnswerBackbone LLM=GPT-5-mini2026.0235.826.5
REMINDRAGMemorization turns=02026.0335.330.8
LinearRAGBackbone LLM=GPT-4o-mini2026.0234.826.3
GFM-RAGMemorization turns=02026.0334.228.7
MA-RAGBackbone LLM=GPT-4o-mini2026.0234.127.4
RAGentABackbone LLM=GPT-4o-mini2026.0232.229.9
PoGMemorization turns=02026.0331.629.2
all-mpnet-base-v2Memorization turns=02026.0329.626.8
nvidia/NV-Embed-v2Memorization turns=02026.0328.926.1
FaithfulRAGBackbone LLM=GPT-4o-mini2026.0228.822.6
GritLM / GritLM-7BMemorization turns=02026.0328.525.8
GraphRAGBackbone LLM=GPT-4o-mini2026.0226.420.8
HippoRAG2Memorization turns=02026.0324.827.3
DyG-RAGMemorization turns=02026.0320.417.8
Direct AnswerBackbone LLM=GPT-4o-mini2026.0218.313.9
Chain-of-Thought (CoT)Memorization turns=02026.0315.813.6
Direct AnsweringMemorization turns=02026.0314.913.2