List your company

Multi-hop Question Answering on Average (test)

51.3EMARBOR
14.27623.88833.543.112Jun 2, 2026

Evaluation results

MethodLinks
ARBORBackbone=Qwen3-14B2026.0651.36062.8
GRPOBackbone=Qwen3-14B2026.0649.85960.8
DAPOBackbone=Qwen3-14B2026.0649.25859.1
ARBORBackbone=Qwen3-8B2026.0648.957.961.1
ARBORBackbone=Qwen3-4B2026.0647.956.759.6
DAPOBackbone=Qwen3-8B2026.0647.155.156.7
GRPOBackbone=Qwen3-8B2026.0646.955.556.9
GRPOBackbone=Qwen3-4B2026.0646.154.955.6
DAPOBackbone=Qwen3-4B2026.0646.155.156.1
TIR PromptingBackbone=Qwen3-4B2026.0629.738.143.6
TIR PromptingBackbone=Qwen3-14B2026.0626.233.235.2
TIR PromptingBackbone=Qwen3-8B2026.0615.720.522.7