Beyond Quantity: Trajectory Diversity Scaling for Code Agents
Original article: https://arxiv.org/abs/2602.03219v2
As code large language models (LLMs) evolve into tool-interactive agents via the Model Context Protocol (MCP), their generalization is increasingly limited by low-quality synthetic data and the diminishing returns of qua...
This entry is part of the Top 50 AI Agent Articles curation.
