List your company
ResearchFormerly WizWand

Knowledge-intensive Reasoning on Musique

87AccuracyR1-Searcher
-1.421.5544.567.45Jan 30, 2026Feb 20, 2026Mar 14, 2026Apr 5, 2026Apr 27, 2026May 19, 2026Jun 10, 2026
Updated 4mo ago

Evaluation results

MethodLinks
R1-SearcherCategory=RL-only TIR Methods2026.0187
AutoTrajCategory=SFT-RL TIR Methods2026.0186
AutoTIRCategory=RL-only TIR Methods2026.0185
Qwen2.5-7B-InstructFramework=Multi-Dimensional Trajectory Evaluation2026.0177
Vanilla SFT-RL TIRCategory=SFT-RL TIR Methods2026.0177
ReSearchCategory=RL-only TIR Methods2026.0176
Tool-Star-SFTCategory=SFT-only TIR Methods2026.0173
ToRLCategory=RL-only TIR Methods2026.0172
Tool-StarCategory=SFT-RL TIR Methods2026.0163
APPOBackbone=Llama3.1-8B-Instruct2026.0635.2
ARPOBackbone=Llama3.1-8B-Instruct2026.0634.8
TEPOdenseBackbone=Llama3.1-8B-Instruct2026.0233.4
TEPOdenseBackbone=Qwen2.5-7B-Instruct2026.0232.2
CISPOBackbone=Llama3.1-8B-Instruct2026.0632.2
GRPOBackbone=Llama3.1-8B-Instruct2026.0232.1
GIGPOBackbone=Llama3.1-8B-Instruct2026.0631.8
GRPOBackbone=Qwen2.5-7B-Instruct2026.0231.7
GPPOBackbone=Qwen2.5-7B-Instruct2026.0631.5
APPOBackbone=Qwen2.5-7B-Instruct2026.0631.5
GIGPOBackbone=Qwen2.5-7B-Instruct2026.0631.1
ARPOBackbone=Qwen2.5-7B-Instruct2026.0631.1
AEPOBackbone=Qwen2.5-7B-Instruct2026.0231
GRPOBackbone=Llama3.1-8B-Instruct2026.0631
GRPOBackbone=Qwen2.5-7B-Instruct2026.0630.6
SFTBackbone=Qwen2.5-7B-Instruct2026.0230.2
TEPOsparseBackbone=Llama3.1-8B-Instruct2026.0230.2
TEPOsparseBackbone=Qwen2.5-7B-Instruct2026.0229.9
Reinforce++Backbone=Llama3.1-8B-Instruct2026.0629.9
ARPOBackbone=Qwen2.5-7B-Instruct2026.0229.83
GPPOBackbone=Llama3.1-8B-Instruct2026.0629.8
ARPOBackbone=Llama3.1-8B-Instruct2026.0229.2
DAPOBackbone=Llama3.1-8B-Instruct2026.0629.2
CISPOBackbone=Qwen2.5-7B-Instruct2026.0629.1
SFTBackbone=Llama3.1-8B-Instruct2026.0229
DAPOBackbone=Qwen2.5-7B-Instruct2026.0628.6
Reinforce++Backbone=Qwen2.5-7B-Instruct2026.0625.2
Qwen2.5-7B-InstructBackbone=Qwen2.5-7B-Instruct2026.0219.2
Llama3.1-8B-InstructBackbone=Llama3.1-8B-Instruct2026.0217.3
AutoTIRBackbone=Qwen2.5-7B, Training Paradigm=RL-only2026.0117
TIR PromptingBackbone=Llama3.1-8B-Instruct2026.0615.5
AutoTrajBackbone=Qwen2.5-7B, Training Paradigm=SFT–RL TIR, SFT Trajectory Count=13K2026.0114.5
ReSearchBackbone=Qwen2.5-7B, Training Paradigm=RL-only2026.0114
Vanilla SFT-RL TIRBackbone=Qwen2.5-7B, Training Paradigm=SFT–RL TIR2026.0114
Tool-Star-SFTBackbone=Qwen2.5-7B, Training Paradigm=SFT-only, SFT Trajectory Count=54K2026.0111
Qwen2.5-7B-InstructBackbone=Qwen2.5-7B, Training Paradigm=Instruct2026.0110.5
Zero-shotBackbone=Llama3.1-8B-Instruct2026.0610.4
TIR PromptingBackbone=Qwen2.5-7B-Instruct2026.069.5
Tool-StarBackbone=Qwen2.5-7B, Training Paradigm=SFT–RL TIR, SFT Trajectory Count=54K2026.019
R1-SearcherBackbone=Qwen2.5-7B, Training Paradigm=RL-only2026.018
Zero-shotBackbone=Qwen2.5-7B-Instruct2026.066.6
ToRLBackbone=Qwen2.5-7B, Training Paradigm=RL-only2026.012