List your company

Interactive Agent on WebShop (Pass@1, Pass@128)

68.65Pass@1ACE
-2.74615.789534.32552.8605Mar 17, 2026

Evaluation results

MethodLinks
ACEBackbone=Qwen2.5-7B2026.0368.6586.8
GRPO-RLVRBackbone=Qwen2.5-7B2026.0367.4585.4
LEAFEBackbone=Qwen2.5-7B2026.0366.587.8
EARLYEXPBackbone=Qwen2.5-7B2026.0361.5584.6
LEAFEBackbone=Llama3.1-8B2026.0356.2581
GRPO-RLVRBackbone=Llama3.1-8B2026.0354.9579.4
ACEBackbone=Llama3.1-8B2026.0354.3579.8
EARLYEXPBackbone=Llama3.1-8B2026.0351.1377.8
BASE (NO FT)Backbone=Qwen2.5-7B2026.030.055.2
BASE (NO FT)Backbone=Llama3.1-8B2026.0301.8