Can Large Language Models Keep Up? Benchmarking Online Adaptation to Continual Knowledge Streams

Jiyeon Kim, Hyunji Lee, Dylan Zhou +6

LLMs operating in dynamic real-world contexts often encounter knowledge that evolves continuously or emerges incrementally. To remain accurate and effective, models must adapt to n…