Empathetic Dialogue on Sentient Benchmark
82.4ScoreGemini2.5-Pro-0605
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Gemini2.5-Pro-06052025.07 | 82.4 | 55 | 4 | |
| GPT-4o-03262025.07 | 79.9 | 51 | 4 | |
| RLVER (PPO)Think=Yes2025.07 | 79.2 | 42 | 9 | |
| RLVER (GRPO)Think=Yes2025.07 | 72 | 34 | 10 | |
| RLVER (GRPO)Think=No2025.07 | 68.3 | 26 | 10 | |
| GPT-4.1-04142025.07 | 68.2 | 35 | 13 | |
| Gemini2.5-Flash-Think-05202025.07 | 66.1 | 39 | 14 | |
| OpenAI-o3-04162025.07 | 62.7 | 32 | 14 | |
| RLVER (PPO)Think=No2025.07 | 61.7 | 24 | 23 | |
| Qwen2.5-7B-Instruct2025.07 | 13.3 | 2 | 76 |