UltraIF: Advancing Instruction Following from the Wild
This addresses the challenge of improving instruction-following in LLMs for broader AI assistant applications, though it appears incremental as it builds on existing models and methods.
The paper tackles the problem of bridging the gap between open-source and leading company LLMs in following complex instructions by proposing UltraIF, a simple and scalable approach that aligns LLaMA-3.1-8B-Base to catch up with its instruct version on 5 benchmarks without benchmark information, using only 8B models for generation and evaluation.
Instruction-following made modern large language models (LLMs) helpful assistants. However, the key to taming LLMs on complex instructions remains mysterious, for that there are huge gaps between models trained by open-source community and those trained by leading companies. To bridge the gap, we propose a simple and scalable approach UltraIF for building LLMs that can follow complex instructions with open-source data. UltraIF first decomposes real-world user prompts into simpler queries, constraints, and corresponding evaluation questions for the constraints. Then, we train an UltraComposer to compose constraint-associated prompts with evaluation questions. This prompt composer allows us to synthesize complicated instructions as well as filter responses with evaluation questions. In our experiment, for the first time, we successfully align LLaMA-3.1-8B-Base to catch up with its instruct version on 5 instruction-following benchmarks without any benchmark information, using only 8B model as response generator and evaluator. The aligned model also achieved competitive scores on other benchmarks. Moreover, we also show that UltraIF could further improve LLaMA-3.1-8B-Instruct through self-alignment, motivating broader use cases for the method. Our code is available at https://github.com/kkk-an/UltraIF.