Performance Comparison of CodeLlama and StarCoder on HumanEval Pro with Multi-Step Java Generation Tasks

We introduce self-invoking code generation, a new task designed to evaluate the progressive reasoning and problem-solving capabilities of LLMs. In this task, models are presented with a base problem and a related, more complex problem. They must solve the base problem and then utilize its solution to address the more complex one. This work features three key contributions. First, we propose a general recipe for generating more challenging versions of existing benchmarks, resulting in three new benchmarks: HumanEval Pro, MBPP Pro, and BigCodeBench-Lite Pro, specifically designed to assess LLMs

Research goal: How does the pass@1 performance of CodeLlama and StarCoder compare on HumanEval Pro when evaluated with multi-step self-invoking Java code generation tasks, and what is the correlation with their performance on the original HumanEval benchmark?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 7.6/10.