User Action Distribution Prediction on Mental Health Support User Action Distributions 160K
0.1586BCE Loss (Total)UserSim-160K-it2
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| UserSim-160K-it2Evaluation Protocol=Fine-tuned, Adversarial Training Iteration=22025.12 | 0.1586 | 0.0597 | 0.137 | 0.104 | 0.2807 | |
| UserSim-160K-baseEvaluation Protocol=Fine-tuned, Adversarial Training Iteration=Base (0)2025.12 | 0.1587 | 0.0599 | 0.1368 | 0.1047 | 0.2806 | |
| UserSim-160K-it3Evaluation Protocol=Fine-tuned, Adversarial Training Iteration=32025.12 | 0.1587 | 0.0597 | 0.1366 | 0.1041 | 0.2811 | |
| UserSim-160K-it1Evaluation Protocol=Fine-tuned, Adversarial Training Iteration=12025.12 | 0.1596 | 0.0598 | 0.1373 | 0.1043 | 0.2833 | |
| GPT-4oEvaluation Protocol=Zero-shot2025.12 | 0.1866 | 0.0877 | 0.1857 | 0.1064 | 0.2952 | |
| Kimi-K2-InstructEvaluation Protocol=Zero-shot2025.12 | 0.1912 | 0.0645 | 0.2063 | 0.1041 | 0.3095 | |
| Llama-3.3-70B-InstructEvaluation Protocol=Zero-shot2025.12 | 0.2072 | 0.0627 | 0.2455 | 0.1047 | 0.3221 |