1 paper · 1 filter
Rahul Suresh Babu, Shashank Indukuri
Tool-augmented language-model agents are often evaluated by whether they select the correct tool, produce valid API arguments, and complete the requested task. However, an agent ma…