Adaptively expanding and compressing prompt-scaffold guidance during RL lets LLM agents train to competitive performance without any skill library at test time.
arXiv preprint arXiv:2606.15390 , year =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
PATS: Policy-Aware Training Scaffolding for Agentic Reinforcement Learning
Adaptively expanding and compressing prompt-scaffold guidance during RL lets LLM agents train to competitive performance without any skill library at test time.