2024
M4LE: A Multi-Ability Multi-Range Multi-Task Multi-Domain Long-Context Evaluation Benchmark for Large Language Models
ACL 2024long
Managing long sequences has become an important and necessary feature for large language models (LLMs). However, assessing their ability to handle long contexts remains a challenge. This paper introduces M4LE, a Multi-ability, Multi-range, Multi-task, Multi-domain benchmark for Long-context Evaluati…