AI RESEARCH
Tool Calling is Linearly Readable and Steerable in Language Models
arXiv CS.AI
•
ArXi:2605.07990v1 Announce Type: cross When a tool-calling agent picks the wrong tool, the failure is invisible until execution: the email gets sent, the meeting gets missed. Probing 12 instruction-tuned models across Gemma 3, Qwen 3, Qwen 2.5, and Llama 3.1 (270M to 27B), we find the identity of the chosen tool is linearly readable and steerable inside the model.