List your company

Multi-hop Question Answering on HotpotQA fullwiki (val)

58.3Exact Match (EM)SDP
29.169636.732344.29551.8577May 12, 2026May 21, 2026May 30, 2026Jun 9, 2026Jun 18, 2026Jun 27, 2026Jul 7, 2026

Evaluation results

MethodLinks
SDPLLM=GPT-4o2026.0558.367.2—
PRISMLLM=GPT-4o2026.0554.267—
IRCoTLLM=GPT-32026.0549.360.7—
DynaKRAGBackbone=GPT-4o-mini2026.0749.2262.181,588.5
DynaKRAGBackbone=Qwen2.5-7B2026.0747.3259.982,373.1
S2G-RAGBackbone=GPT-4o-mini2026.0744.9860.951,766
CoRAG-s3Backbone=GPT-4o-mini2026.0744.9661.082,569.7
IRCoTBackbone=GPT-4o-mini2026.0744.8560.831,545.7
S2G-RAGBackbone=Qwen2.5-7B2026.0744.2557.12,807.3
CRAGBackbone=GPT-4o-mini2026.0744.0659.511,355.4
Self-Ask+SearchBackbone=GPT-4o-mini2026.0744.0460.233,432.1
IRCoTBackbone=Qwen2.5-7B2026.0743.3956.171,723.5
CoRAG-s3Backbone=Qwen2.5-7B2026.0742.8255.732,814.3
CRAGBackbone=Qwen2.5-7B2026.0741.8154.451,720.1
Fixed-KBackbone=GPT-4o-mini2026.0741.757.811,745.1
Self-Ask+SearchBackbone=Qwen2.5-7B2026.0740.4352.733,302
Fixed-KBackbone=Qwen2.5-7B2026.0739.5851.771,892.3
SETR-CoT & IRILLM=GPT-4o2026.0539.240.5—
PAR2-RAGBackbone=GPT-4o-mini2026.0739.1855.022,776
Adaptive-RAGBackbone=Qwen2.5-7B2026.0738.6250.421,596.4
PAR2-RAGBackbone=Qwen2.5-7B2026.0737.7949.523,053.6
DynaKRAGBackbone=Llama-3.1-8B2026.0737.4248.761,868.6
CRAGBackbone=Llama-3.1-8B2026.0735.6447.7968.8
S2G-RAGBackbone=Llama-3.1-8B2026.0735.1747.612,843
RankZephyrLLM=GPT-4o2026.0534.735—
RankGPTLLM=GPT-4o2026.0534.635.3—
Adaptive-RAGBackbone=GPT-4o-mini2026.0734.1947.831,120.4
RankZephyr + CoTLLM=GPT-4o2026.053434.4—
CoRAG-s3Backbone=Llama-3.1-8B2026.0732.8444.992,576.3
IRCoTBackbone=Llama-3.1-8B2026.0732.5244.661,657.1
Adaptive-RAGBackbone=Llama-3.1-8B2026.0732.1543.951,197.2
Self-Ask+SearchBackbone=Llama-3.1-8B2026.0732.1143.543,525.1
PAR2-RAGBackbone=Llama-3.1-8B2026.0731.5643.213,035.5
Fixed-KBackbone=Llama-3.1-8B2026.0730.2943.111,799.4