List your company

Action Prediction on Human Evaluation User Actions Dataset (test)

79Win RateLongNAP
27.5240.88554.2567.615Mar 6, 2026

Evaluation results

MethodLinks
LongNAPModel=Qwen 2.5 VL2026.0379
Few-shot RAGModel=Gemini2026.0358.5
Zero-shotModel=Gemini2026.0355
Few-shot RAGModel=Qwen 2.5 VL2026.0345
Zero-shotModel=Qwen 2.5 VL2026.0333
SFTModel=Qwen 2.5 VL2026.0329.5