cs.AINov 6, 2026

META-IR: Meta-Introspective Reasoning for Adaptive Safety Alignment in LLM

Authors: Mahabubur Rahman Miraj, Hongyu Huang, Nankun Mu, Chao Chen, Shanqin Yi

Organizations: College of Computer Science, Chongqing University, Chongqing, China · School of Economics and Business Administration, Chongqing University, Chongqing, China

Abstract

Large Language Models (LLMs) remain vulnerable to jailbreak attacks and harmful-content prompts, particularly under adversarial distribution shifts where static alignment pipelines fail to generalize. Although introspective methods such as STAIR strengthen refusal consistency, they depend on fixed self-evaluation policies that cannot rapidly adapt to evolving manipulation strategies. To overcome these limitations, this paper introduces META-IR, a meta-introspective reasoning framework that enhances safety alignment through meta-learning-driven adaptation. It employs a dual-loop architecture: an inner introspective loop that performs structured risk analysis, self-assessment, and reflective correction via SI-MCTS, and an outer meta-learning loop that optimizes the introspection policy across diverse jailbreak task families to enable rapid generalization to unseen attacks. A dynamic jailbreak simulation module further exposes the model to progressively harder adversarial prompts, enabling continual self-improvement without human supervision. Experiments across multiple models (Llama-3.1-8B-Instruct and Qwen-2-7B-Instruct) and a broad suite of safety, helpfulness, generalization, and introspection metrics demonstrate that META-IR substantially improves adaptation speed, robustness, and reasoning stability compared with all baseline models. These results establish META-IR as an effective pathway toward building LLMs capable of sustained, adaptive safety alignment under dynamic and evolving threat environments.

Figures & tables

Explore similar work

CardsList
  1. Metis: Learning to Jailbreak LLMs via Self-Evolving Metacognitive Policy Optimization

    May 11, 2026Huilin Zhou, Jian Zhao, Yilu Zhong +7Inference-Time OptimizationAdversarial Attacks on LLMs

  2. Beyond Refusal Patterns: Safe-Role Internalization for Robust and Generalizable LLM Safety Alignment

    Oct 4, 2026Jinghao Pang, Jitai Hao, Qiang Huang +2LLM AlignmentLLM Safety Alignment

  3. MENTOR: A Metacognition-Driven Self-Evolution Framework for Uncovering and Mitigating Implicit Domain Risks in LLMs

    Nov 10, 2025Liang Shan, Kaicheng Shen, Wen Wu +8LLM SafetyLLM Safety Alignment