Human-Agent Alignment Evaluation on TUX dataset
68.83Matched RateLlama-3.1-8B
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Llama-3.1-8BTrait space=active 8-trait space, Matching Strategy=Nearest neighbors2026.05 | 68.83 | 67.76 | 1.06 | 1.5699 | 3.8601 | 0.2487 | |
| Mistral-NemoTrait space=active 8-trait space, Matching Strategy=Nearest neighbors2026.05 | 67.23 | 66.67 | 0.56 | 0.8453 | 1.6835 | 0.1084 | |
| Qwen2.5-7BTrait space=active 8-trait space, Matching Strategy=Nearest neighbors2026.05 | 67.07 | 66.77 | 0.3 | 0.4522 | 1.7967 | 0.1157 | |
| Qwen3-14BTrait space=active 8-trait space, Matching Strategy=Nearest neighbors2026.05 | 66.99 | 65.77 | 1.22 | 1.8575 | 4.5498 | 0.2931 |