Character Relationship Network Updating on HoLLMwood Synthesized scenario
85Win RateGemini-2.5-Pro
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Gemini-2.5-ProBaseline=vs. HoLLMwood2026.04 | 85 | 4.5 | |
| Qwen3-32BBaseline=vs. HoLLMwood2026.04 | 85 | 3.18 | |
| Llama3.3-70BBaseline=vs. HoLLMwood2026.04 | 80 | 3.49 | |
| DeepSeek-V3.2-ThinkBaseline=vs. HoLLMwood2026.04 | 65 | 4.08 | |
| DeepSeek-V3.2Baseline=vs. HoLLMwood2026.04 | 63.3 | 4.08 |