online-augmentation

This paper investigates whether online skill and memory modules for web agents are worth their token cost under a fixed inference budget, finding that a budget-matched vanilla baseline often matches or outperforms augmented methods across three domains and models.