List your company

Web-based Agent Interaction on WebShop (test)

73Success RateFed-SE
-2.9216.7936.556.21Dec 9, 2025Dec 29, 2025Jan 19, 2026Feb 9, 2026Mar 2, 2026Mar 23, 2026Apr 13, 2026

Evaluation results

MethodLinks
Fed-SEBase Model=Qwen2.5-7B2025.1273
CentralizedBase Model=Qwen2.5-7B2025.1271
Fed-SEBase Model=Qwen3-8B2025.1268
FedITBase Model=Qwen2.5-7B2025.1267.5
Fed-SEBase Model=Llama2-7B2025.1266
LocalBase Model=Qwen2.5-7B2025.1265
Fed-SEBase Model=Qwen2.5-3B2025.1265
LocalBase Model=Qwen3-8B2025.1264.5
FedITBase Model=Qwen3-8B2025.1264.5
FedITBase Model=Qwen2.5-3B2025.1263.5
CentralizedBase Model=Qwen2.5-3B2025.1261.5
FedITBase Model=Llama2-7B2025.1261
LocalBase Model=Qwen2.5-3B2025.1260
CentralizedBase Model=Llama2-7B2025.1259
Fed-SEBase Model=Qwen3-1.7B2025.1257.5
LocalBase Model=Llama2-7B2025.1254
LocalBase Model=Qwen3-1.7B2025.1252
FedITBase Model=Qwen3-1.7B2025.1250.5
CentralizedBase Model=Qwen3-1.7B2025.1244.5
CentralizedBase Model=Qwen3-8B2025.1233.5
GPT-4oBackbone=GPT-4o2026.0431.39
GPT-4o-miniBackbone=GPT-4o-mini2026.0425.3
Qwen2-7B + MISEBackbone=Qwen2-7B-Instruct, Technique=MISE2026.0425.15
LLaMA3-8B + MISEBackbone=LLaMA3-8B-Instruct, Technique=MISE2026.0424.34
Qwen2-7B + PRMBackbone=Qwen2-7B-Instruct, Technique=PRM2026.0423.81
Qwen2-7B + PPOBackbone=Qwen2-7B-Instruct, Technique=PPO2026.0423.75
Qwen2-7BBackbone=Qwen2-7B-Instruct2026.0423.38
LLaMA3-8B + PRMBackbone=LLaMA3-8B-Instruct, Technique=PRM2026.0422.81
LLaMA3-8B + PPOBackbone=LLaMA3-8B-Instruct, Technique=PPO2026.0422.47
LLaMA3-8B + RFTBackbone=LLaMA3-8B-Instruct, Technique=RFT2026.0421.97
LLaMA3-8B + online DPOBackbone=LLaMA3-8B-Instruct, Technique=online DPO2026.0420.96
LLaMA3-8BBackbone=LLaMA3-8B-Instruct2026.0420.14
LLaMA3-8B + ReActBackbone=LLaMA3-8B-Instruct, Technique=ReAct2026.0419.98
Pre-TrainedBase Model=Qwen3-1.7B2025.127
Pre-TrainedBase Model=Qwen2.5-3B2025.123
Pre-TrainedBase Model=Qwen2.5-7B2025.122
Pre-TrainedBase Model=Qwen3-8B2025.121.5
Pre-TrainedBase Model=Llama2-7B2025.120
Gemma2-9BBackbone=Gemma2-9B-Instruct2026.040
Gemma2-9B + PPOBackbone=Gemma2-9B-Instruct, Technique=PPO2026.040
Gemma2-9B + PRMBackbone=Gemma2-9B-Instruct, Technique=PRM2026.040
Gemma2-9B + MISEBackbone=Gemma2-9B-Instruct, Technique=MISE2026.040