Evaluating Long Range Dependency Handling in Code Generation LLMs

Yannick Assogba, Donghao Ren

As language models support larger and larger context sizes, evaluating their ability to make effective use of that context becomes increasingly important. We analyze the ability of…