SEQUOR

New benchmarks reveal LLMs struggle with multi-turn instruction following

Researchers have introduced two new benchmarks to evaluate large language models' ability to follow complex instructions. SEQUOR addresses constraint adherence in long, multi-turn conversations, revealing that model acc…