A Dual-Phase Self-Evolution Framework for Large Language Models

Researchers at the Association for Computational Linguistics introduced a Dual-Phase Self-Evolution (DPSE) framework for large language models that jointly optimizes user preference adaptation and domain-specific competence. The framework uses a Censor module to extract interaction signals and guide data expansion, outperforming existing methods on general NLP benchmarks and long-term dialogue tasks. This work provides an autonomous path for continual self-evolution of LLMs.

Abstract The capabilities of Large Language Models LLMs are limited to some extent by pre-training, so some researchers optimize LLMs through post-training. Existing post-training strategies, such as memory-based retrieval or preference optimization, improve user alignment yet fail to enhance the model’s domain cognition. To bridge this gap, we propose a novel Dual-Phase Self-Evolution DPSE framework that jointly optimizes user preference adaptation and domain-specific competence. DPSE introduces a Censor module to extract multi-dimensional interaction signals and estimate satisfaction scores, which guide structured data expansion via topic-aware and preference-driven strategies. These expanded datasets support a two-stage fine-tuning pipeline: supervised domain grounding followed by frequency-aware preference optimization. Experiments across general NLP benchmarks and long-term dialogue tasks demonstrate that DPSE consistently outperforms Supervised Fine-Tuning, Preference Optimization, and Memory-Augmented baselines. Ablation studies validate the contribution of each module. In this way, our framework provides an autonomous path toward continual self-evolution of LLMs.- Anthology ID: - 2026.findings-acl.37 - Volume: Findings of the Association for Computational Linguistics: ACL 2026 /volumes/2026.findings-acl/ - Month: - July - Year: - 2026 - Address: - San Diego, California, United States - Editors: Maria Liakata /people/maria-liakata/ , Viviane P. Moreira /people/viviane-p-moreira/unverified/ , Jiajun Zhang /people/jiajun-zhang/unverified/ , David Jurgens /people/david-jurgens/ - Venue: Findings /venues/findings/ - SIG: - Publisher: - Association for Computational Linguistics - Note: - Pages: - 772–782 - Language: - URL: https://aclanthology.org/2026.findings-acl.37/ https://aclanthology.org/2026.findings-acl.37/ - DOI: - Cite ACL : - Haoran Sun, Zekun Zhang, and Shaoning Zeng. 2026. A Dual-Phase Self-Evolution Framework for Large Language Models https://aclanthology.org/2026.findings-acl.37/ . In Findings of the Association for Computational Linguistics: ACL 2026 , pages 772–782, San Diego, California, United States. Association for Computational Linguistics. - Cite Informal : A Dual-Phase Self-Evolution Framework for Large Language Models https://aclanthology.org/2026.findings-acl.37/ Sun et al., Findings 2026 - PDF: https://aclanthology.org/2026.findings-acl.37.pdf https://aclanthology.org/2026.findings-acl.37.pdf