AdvancedIF: Rubric-Based Benchmarking and Reinforcement Learning for Advancing LLM Instruction Following

TL;DR

This paper introduces AdvancedIF, a new benchmark with expert rubrics for evaluating complex instruction following in LLMs, and RIFL, a reinforcement learning pipeline that significantly improves LLM performance on this benchmark.

Contribution

The paper presents a comprehensive rubric-based benchmark and a novel reinforcement learning method, RIFL, to enhance LLM instruction-following capabilities for complex tasks.

Findings

RIFL improves LLM performance by 6.7% on AdvancedIF

Rubrics effectively guide training and evaluation of instruction following

RIFL achieves strong results on public benchmarks

Abstract

Recent progress in large language models (LLMs) has led to impressive performance on a range of tasks, yet advanced instruction following (IF)-especially for complex, multi-turn, and system-prompted instructions-remains a significant challenge. Rigorous evaluation and effective training for such capabilities are hindered by the lack of high-quality, human-annotated benchmarks and reliable, interpretable reward signals. In this work, we introduce AdvancedIF (we will release this benchmark soon), a comprehensive benchmark featuring over 1,600 prompts and expert-curated rubrics that assess LLMs ability to follow complex, multi-turn, and system-level instructions. We further propose RIFL (Rubric-based Instruction-Following Learning), a novel post-training pipeline that leverages rubric generation, a finetuned rubric verifier, and reward shaping to enable effective reinforcement learning for instruction following. Extensive experiments demonstrate that RIFL substantially improves the instruction-following abilities of LLMs, achieving a 6.7% absolute gain on AdvancedIF and strong results on public benchmarks. Our ablation studies confirm the effectiveness of each component in RIFL. This work establishes rubrics as a powerful tool for both training and evaluating advanced IF in LLMs, paving the way for more capable and reliable AI systems.