List your company

Multi-Turn User Simulation on PEARL (test)

95.31Success Rate (SR)UserSim-Qwen 8B
33.783649.756865.7381.7032Mar 19, 2026

Evaluation results

MethodLinks
UserSim-Qwen 8BMode=Fine-tuned, Parameters=8B2026.0395.310.02322.37
UserSim-Llama 8BMode=Fine-tuned, Parameters=8B2026.0392.650.04982.37
Qwen3 32BMode=Zero-shot, Parameters=32B2026.0377.720.009721.31
Llama3.1 70BMode=Zero-shot, Parameters=70B2026.0336.150.003363.51