List your company

Language Agent Task on ALFWorld (OOD)

58.7ScoreTask-Perturbed NLL Optimization
55.0656.00556.9557.895Jun 25, 2026

Evaluation results

MethodLinks
Task-Perturbed NLL OptimizationBackbone=Qwen3-8B2026.0658.7
COINBackbone=Qwen3-8B2026.0657.3
Task AugBackbone=Qwen3-8B2026.0656.3
SFTBackbone=Qwen3-8B2026.0655.2