List your company

Function Selection on TaskBench Multimedia

82.3Function Selection AccRaw Demonstrations
2.2223.0143.864.59Dec 18, 2025

Evaluation results

MethodLinks
Raw DemonstrationsBackbone=Qwen3 14B2025.1282.3
Raw DemonstrationsBackbone=GPT-4o2025.1282.3
No ICLBackbone=Qwen3 14B2025.1281.7
No ICLBackbone=GPT-4o2025.1281.6
Raw DemonstrationsBackbone=Qwen3 8B2025.1280.4
No ICLBackbone=Qwen3 8B2025.1278.5
DTDR-CBackbone=GPT-4o2025.1274.8
DTDR-CBackbone=Qwen3 14B2025.1274
Dependency RetrievalBackbone=Qwen3 14B2025.1273.8
DTDR-LBackbone=GPT-4o2025.1273.7
DTDR-LBackbone=Qwen3 14B2025.1273.4
Dependency RetrievalBackbone=GPT-4o2025.1272.8
Dependency RetrievalBackbone=Qwen3 8B2025.1269.1
DTDR-CBackbone=Qwen3 8B2025.1269.1
DTDR-LBackbone=Qwen3 8B2025.1267.6
DTDR-LBackbone=Qwen3 4B2025.1264.1
Learned RetrievalBackbone=Qwen3 4B2025.1258.2
Raw DemonstrationsBackbone=Qwen3 4B2025.1257.3
No ICLBackbone=Qwen3 4B2025.1254.3
DTDR-CBackbone=Qwen3 4B2025.1252.4
Dependency RetrievalBackbone=Qwen3 4B2025.1251.8
Learned RetrievalBackbone=GPT-4o2025.1251.1
Learned RetrievalBackbone=Qwen3 14B2025.1250.8
Learned RetrievalBackbone=Qwen3 8B2025.1249.6
DTDR-CBackbone=Qwen3 1.7B2025.1247.8
Raw DemonstrationsBackbone=Qwen3 1.7B2025.1246.9
DTDR-LBackbone=Qwen3 1.7B2025.1245.5
Dependency RetrievalBackbone=Qwen3 1.7B2025.1245.1
Learned RetrievalBackbone=Qwen3 1.7B2025.1242.1
No ICLBackbone=Qwen3 1.7B2025.1238.1
DTDR-CBackbone=Qwen3 0.6B2025.1220.2
DTDR-LBackbone=Qwen3 0.6B2025.1217.6
Learned RetrievalBackbone=Qwen3 0.6B2025.1212.1
Dependency RetrievalBackbone=Qwen3 0.6B2025.129.5
No ICLBackbone=Qwen3 0.6B2025.127.3
Raw DemonstrationsBackbone=Qwen3 0.6B2025.125.3