List your company

Next-action prediction on OPeRA (test)

52.92Action Generation AccAlignUSER+
4.549617.107329.66542.2227Jul 23, 2025Aug 23, 2025Sep 23, 2025Oct 24, 2025Nov 24, 2025Dec 25, 2025Jan 26, 2026

Evaluation results

MethodLinks
AlignUSER+2026.0152.9271.9466.8880.52——
AlignUSER2026.0151.4769.8166.2978.07——
Shop-R1 (Qwen2.5-VL-3B-Instruct)Settings=SFT + RL2025.0738.44———57.2757.69
ContextSim2026.0135.3764.2261.0373.74——
SimUSER2026.0124.2152.4448.6859.63——
SimUSER2026.0124.2152.4448.6859.63——
Agent4Rec2026.0123.0950.0546.3756.7——
Agent4Rec2026.0123.0950.0546.3756.7——
RecAgent2026.0122.7149.1845.2554.12——
RecAgent2026.0122.7149.1845.2554.12——
GPT-4.1persona=excluded, rationale=included2026.0122.0645.5543.4558.47——
GPT-4.1Ablation=w/o persona2026.0122.0645.5543.4558.47——
GPT-4.1persona=included, rationale=included2026.0121.5148.7844.4747.54——
GPT-4.1Ablation=Full2026.0121.5148.7844.4747.54——
GPT-4.1persona=included, rationale=excluded2026.0121.2834.9342.6351.17——
GPT-4.1Ablation=w/o rationale2026.0121.2834.9342.6351.17——
Qwen2.5-VL-3B-InstructSettings=SFT2025.0720.23———60.8653.95
Claude-3.7-Sonnetpersona=included, rationale=included2026.0110.7531.5827.2743.52——
Claude-3.7-Sonnetpersona=excluded, rationale=included2026.0110.7525.3322.7643.1——
Claude-3.7Ablation=Full, Model Detail=Claude-3.7-Sonnet2026.0110.7531.5827.2743.52——
Claude-3.7Ablation=w/o persona, Model Detail=Claude-3.7-Sonnet2026.0110.7525.3322.7643.1——
Claude-3.7-Sonnetpersona=included, rationale=excluded2026.0110.0826.0620.2943.1——
Claude-3.7Ablation=w/o rationale, Model Detail=Claude-3.7-Sonnet2026.0110.0826.0620.2943.1——
Llama-3.3-70B-Instructpersona=included, rationale=excluded2026.018.7623.619.2234.19——
Llama-3.3Ablation=w/o rationale, Model Detail=Llama-3.3-70B-Instruct2026.018.7623.619.2234.19——
Llama-3.3-70B-Instructpersona=included, rationale=included2026.018.3124.2919.9936.64——
Llama-3.3-70B-Instructpersona=excluded, rationale=included2026.018.3123.6918.5933.21——
Llama-3.3Ablation=Full, Model Detail=Llama-3.3-70B-Instruct2026.018.3124.2919.9936.64——
Llama-3.3Ablation=w/o persona, Model Detail=Llama-3.3-70B-Instruct2026.018.3123.6918.5933.21——
Claude-3.5-SonnetSettings=Zero-shot Prompt2025.077.66———58.8345.61
Qwen2.5-VL-3B-InstructSettings=Zero-shot Prompt2025.076.41———34.4538.79