Narrative Generation Evaluation on Creativityprism 2025 (test)
100Narrative Coherence (NC)BiT-MCTS
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| BiT-MCTSLLM Judge=DeepSeek-V3, Against=Vanilla2026.03 | 100 | 95 | 92.5 | 100 | 95 | 100 | 75 | 75 | 100 | 100 | 93.3 | |
| BiT-MCTSLLM Judge=GPT-5-Mini, Against=Vanilla2026.03 | 100 | 80 | 90 | 90 | 90 | 70 | 70 | 70 | 100 | 100 | 86 | |
| BiT-MCTSLLM Judge=GPT-5-Mini, Against=StoryWriter2026.03 | 100 | 95 | 70 | 95 | 75 | 70 | 60 | 60 | 90 | 95 | 81 | |
| BiT-MCTSLLM Judge=Gemini-2.5-Flash, Against=StoryWriter2026.03 | 85 | 65 | 85 | 75 | 85 | 55 | 75 | 75 | 80 | 80 | 76 | |
| BiT-MCTSLLM Judge=Gemini-2.5-Flash, Against=Vanilla2026.03 | 80 | 80 | 70 | 80 | 80 | 80 | 55 | 55 | 85 | 90 | 75.5 | |
| BiT-MCTSLLM Judge=GPT-5-Mini, Against=Narrative Studio2026.03 | 70 | 70 | 80 | 90 | 90 | 75 | 60 | 60 | 90 | 90 | 77.5 | |
| BiT-MCTSLLM Judge=DeepSeek-V3, Against=StoryWriter2026.03 | 55 | 55 | 75 | 45 | 70 | 55 | 82.5 | 82.5 | 65 | 65 | 65 | |
| BiT-MCTSLLM Judge=Gemini-2.5-Flash, Against=Narrative Studio2026.03 | 55 | 55 | 80 | 75 | 80 | 65 | 80 | 80 | 80 | 80 | 73 | |
| BiT-MCTSLLM Judge=DeepSeek-V3, Against=Narrative Studio2026.03 | 35 | 25 | 65 | 50 | 65 | 70 | 95 | 95 | 85 | 65 | 65 |