List your company

Question Answering on PopQA (F1)

59.9F1 ScoreQwen3-Embedding (8B)
12.99625.17337.3549.527Oct 16, 2025Nov 9, 2025Dec 3, 2025Dec 28, 2025Jan 21, 2026Feb 14, 2026Mar 11, 2026

Evaluation results

MethodLinks
Qwen3-Embedding (8B)Retrieval Category=Dense Retrieval, Reranker=Qwen3-Embedding (8B) + reranker2026.0259.9
Qwen3-Embedding (8B)Retrieval Category=Dense Retrieval, Reranker=None2026.0259.8
PANINIRetrieval Category=Proposed Method2026.0257.56
BM25Retrieval Category=Sparse Retrieval, Reranker=BM25 + reranker2026.0256.8
HippoRAGRetrieval Category=Structure-Augmented RAG2026.0256.2
NV-Embed-v2 (7B)Retrieval Category=Dense Retrieval2026.0255.8
HippoRAG 2Retrieval Category=Structure-Augmented RAG2026.0255.7
RAPTORRetrieval Category=Structure-Augmented RAG2026.0255.1
IRCoTRetrieval Category=Agentic Systems2026.0253.8
IGPOBackbone=Qwen2.5-7B-Instruct2025.1053.8
BM25Retrieval Category=Sparse Retrieval, Reranker=None2026.0253
GraphRAGRetrieval Category=Structure-Augmented RAG2026.0251.3
Search-R1Retrieval Category=Agentic Systems2026.0249.7
GRPOBackbone=Qwen2.5-7B-Instruct2025.1049.6
TIPSBackbone=Qwen-2.5-7B Instruct2026.0349.26
PPOBackbone=Qwen2.5-7B-Instruct2025.1048.7
PPOBackbone=Qwen-2.5-7B Instruct2026.0347.62
TIPSBackbone=Qwen-2.5-3B Instruct2026.0347.43
PPOBackbone=Qwen-2.5-3B Instruct2026.0346.81
MT-PPOBackbone=Qwen-2.5-7B Instruct2026.0346.45
GSPOBackbone=Qwen2.5-7B-Instruct2025.1045.4
MT-PPOBackbone=Qwen-2.5-3B Instruct2026.0344.83
Reinforce++Backbone=Qwen2.5-7B-Instruct2025.1044.3
MT-GRPO*Backbone=Qwen-2.5-7B Instruct2026.0344.11
RLOOBackbone=Qwen2.5-7B-Instruct2025.1043.1
GRPOBackbone=Qwen-2.5-3B Instruct2026.0343
GRPOBackbone=Qwen-2.5-7B Instruct2026.0342.51
MT-GRPO*Backbone=Qwen-2.5-3B Instruct2026.0340.4
NoneRetrieval Category=No Retrieval2026.0222.7
LightRAGRetrieval Category=Structure-Augmented RAG2026.0214.8