List your company

Function Selection on TaskBench DailyLife

96.8Function Selection AccuracyRaw Demonstrations
7.04830.34953.6576.951Dec 18, 2025

Evaluation results

MethodLinks
Raw DemonstrationsBackbone=GPT-4o2025.1296.8
Raw DemonstrationsBackbone=Qwen3 14B2025.1296.6
No ICLBackbone=GPT-4o2025.1296.6
No ICLBackbone=Qwen3 14B2025.1296.2
Raw DemonstrationsBackbone=Qwen3 8B2025.1294.5
No ICLBackbone=Qwen3 8B2025.1293.9
DTDR-LBackbone=Qwen3 14B2025.1293.7
DTDR-LBackbone=GPT-4o2025.1291.5
DTDR-LBackbone=Qwen3 8B2025.1291
DTDR-LBackbone=Qwen3 4B2025.1289
Raw DemonstrationsBackbone=Qwen3 4B2025.1287.6
Dependency RetrievalBackbone=Qwen3 14B2025.1284.1
DTDR-CBackbone=Qwen3 14B2025.1284.1
No ICLBackbone=Qwen3 4B2025.1283.5
DTDR-LBackbone=Qwen3 1.7B2025.1283.1
Dependency RetrievalBackbone=GPT-4o2025.1282.9
DTDR-CBackbone=GPT-4o2025.1282.9
Dependency RetrievalBackbone=Qwen3 8B2025.1275.3
DTDR-CBackbone=Qwen3 8B2025.1275.3
Raw DemonstrationsBackbone=Qwen3 1.7B2025.1269
No ICLBackbone=Qwen3 1.7B2025.1266.7
Dependency RetrievalBackbone=Qwen3 4B2025.1264.7
DTDR-CBackbone=Qwen3 4B2025.1264.7
Learned RetrievalBackbone=Qwen3 4B2025.1260.4
DTDR-CBackbone=Qwen3 1.7B2025.1260.2
Dependency RetrievalBackbone=Qwen3 1.7B2025.1260
Learned RetrievalBackbone=Qwen3 1.7B2025.1256.6
Learned RetrievalBackbone=Qwen3 14B2025.1256.5
Learned RetrievalBackbone=Qwen3 8B2025.1252.2
Learned RetrievalBackbone=GPT-4o2025.1248.1
DTDR-LBackbone=Qwen3 0.6B2025.1245.9
DTDR-CBackbone=Qwen3 0.6B2025.1223.7
Learned RetrievalBackbone=Qwen3 0.6B2025.1219.6
No ICLBackbone=Qwen3 0.6B2025.1212.8
Dependency RetrievalBackbone=Qwen3 0.6B2025.1211.5
Raw DemonstrationsBackbone=Qwen3 0.6B2025.1210.5