Funny caption generation on Electric sheep Low-Humor
79.45Top-1 AccuracyHOMER
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| HOMERBase LLM=GPT-4o2026.02 | 79.45 | 91.48 | 93.81 | |
| HOMERBase LLM=Claude-42026.02 | 71.75 | 89.86 | 95.19 | |
| HumorousAIBase LLM=GPT-4o2026.02 | 69.41 | 80.65 | 85.33 | |
| CLoTBase LLM=GPT-4o2026.02 | 67.49 | 81.16 | 87.83 | |
| Self-consistencyBase LLM=Claude-42026.02 | 67.41 | 82.15 | 87.66 | |
| LoLBase LLM=Claude-42026.02 | 66.23 | 83.04 | 87.66 | |
| HumorousAIBase LLM=Claude-42026.02 | 65.69 | 83.13 | 89.83 | |
| LoLBase LLM=GPT-4o2026.02 | 64.73 | 80.6 | 84.73 | |
| CoTBase LLM=GPT-4o2026.02 | 63.64 | 77.76 | 84.01 | |
| CoTBase LLM=Claude-42026.02 | 63.5 | 74.39 | 78.01 | |
| Self-consistencyBase LLM=GPT-4o2026.02 | 62.02 | 70.47 | 78.78 | |
| Few-shotBase LLM=GPT-4o2026.02 | 55.67 | 72.66 | 83.66 | |
| CLoTBase LLM=Claude-42026.02 | 51.83 | 72.83 | 79.16 | |
| Few-shotBase LLM=Claude-42026.02 | 48.33 | 63.67 | 68.33 | |
| PhunnyBase LLM=Claude-42026.02 | 30.33 | 50.2 | 60.41 | |
| HOMERBase LLM=LLaVA-1.5 (7B)2026.02 | 30.16 | 34.33 | 35.83 | |
| HOMERBase LLM=Qwen-VL (7B)2026.02 | 29.62 | 42.37 | 47.42 | |
| PhunnyBase LLM=GPT-4o2026.02 | 29.31 | 38.32 | 48.05 | |
| Few-shotBase LLM=Qwen-VL (7B)2026.02 | 28.69 | 40.44 | 45.19 | |
| CoTBase LLM=Qwen-VL (7B)2026.02 | 26.83 | 37.39 | 41.83 | |
| LoLBase LLM=Qwen-VL (7B)2026.02 | 24.94 | 38.24 | 44.13 | |
| CLoTBase LLM=LLaVA-1.5 (7B)2026.02 | 24.44 | 27.78 | 28.98 | |
| HumorousAIBase LLM=LLaVA-1.5 (7B)2026.02 | 22.78 | 28.33 | 30.55 | |
| Few-shotBase LLM=LLaVA-1.5 (7B)2026.02 | 20.67 | 25.26 | 27.5 | |
| Self-consistencyBase LLM=Qwen-VL (7B)2026.02 | 19.93 | 28.28 | 33.16 | |
| CLoTBase LLM=Qwen-VL (7B)2026.02 | 18.66 | 33.33 | 41.67 | |
| HumorousAIBase LLM=Qwen-VL (7B)2026.02 | 16.56 | 25.56 | 29.16 | |
| LoLBase LLM=LLaVA-1.5 (7B)2026.02 | 15.24 | 24.76 | 28.57 | |
| PhunnyBase LLM=Qwen-VL (7B)2026.02 | 10.55 | 20.55 | 30.56 | |
| PhunnyBase LLM=LLaVA-1.5 (7B)2026.02 | 10.55 | 20.57 | 27.01 | |
| Self-consistencyBase LLM=LLaVA-1.5 (7B)2026.02 | 7.99 | 10.67 | 13.33 | |
| CoTBase LLM=LLaVA-1.5 (7B)2026.02 | 7.22 | 11.66 | 13.89 |