T^2PO: Uncertainty-Guided Exploration Control for Stable Multi-Turn Agentic Reinforcement Learning
Open paper
Benchmark
Model
Rank
Results
agents-on-alfworld
T2PO (Qwen3-8B-RFT)
#5
Success Rate: 92.41
question-answering-on-natural-questions
T2PO (Qwen3-4B-RFT)
#16
EM: 46.13
Rank counts only results with a code link.