FireBench: Evaluating Instruction Following in Enterprise and API-Driven LLM Applications
arXiv:2603.04857v1 Announce Type: new Abstract: Instruction following is critical for LLMs deployed in enterprise and API-driven settings, where strict adherence to output formats, content constraints, …
Yunfan Zhang, Yijie Bei, Jetashree Ravi, Pawel Garbacki
3 views