ContextGuard: Structured Self-Auditing for Context Learning in Language Models

Hongbo Jin, Chi Wang, Haoran Tang +5

Recent benchmarks reveal that despite strong reasoning capabilities, large language models (LLMs) still struggle to faithfully apply complex contextual knowledge. These failures ar…