Beyond the Final Answer: Evaluating the Reasoning Trajectories of Tool-Augmented Agents

Open in new window