Evaluation of Instruction-Following Ability for Large Language Models on Story-Ending Generation

TL;DR

This paper introduces an automatic evaluation method for assessing how well large language models follow instructions in story-ending generation, showing that open-source models perform nearly as well as GPT-3.5.

Contribution

The paper proposes a novel MRC-based automatic evaluation pipeline for instruction-following ability in story generation tasks, validated against human judgments.

Findings

The proposed metric aligns well with human evaluation.

Open-source LLMs nearly match GPT-3.5 in instruction-following performance.

The evaluation method is effective for diverse, context-specific instructions.

Abstract

Instruction-tuned Large Language Models (LLMs) have achieved remarkable performance across various benchmark tasks. While providing instructions to LLMs for guiding their generations is user-friendly, assessing their instruction-following capabilities is still unclarified due to a lack of evaluation metrics. In this paper, we focus on evaluating the instruction-following ability of LLMs in the context of story-ending generation, which requires diverse and context-specific instructions. We propose an automatic evaluation pipeline that utilizes a machine reading comprehension (MRC) model to determine whether the generated story-ending reflects instruction. Our findings demonstrate that our proposed metric aligns with human evaluation. Furthermore, our experiments confirm that recent open-source LLMs can achieve instruction-following performance close to GPT-3.5, as assessed through automatic evaluation.