![[2602.23320] ParamMem: Augmenting Language Agents with Parametric Reflective Memory](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.23320] ParamMem: Augmenting Language Agents with Parametric Reflective Memory
Summary
The paper introduces ParamMem, a parametric memory module designed to enhance language agents by enabling diverse reflective outputs, improving reasoning performance in tasks like code generation and question answering.
Why It Matters
As language agents become integral in various applications, improving their reasoning capabilities is crucial. ParamMem addresses the limitations of repetitive outputs in self-reflection, potentially leading to more effective AI systems in real-world scenarios.
Key Takeaways
- ParamMem enhances language agents by encoding reflection patterns into model parameters.
- Diverse reflection signals correlate positively with task success, improving reasoning performance.
- The framework supports sample efficiency and weak-to-strong transfer across model scales.
Computer Science > Machine Learning arXiv:2602.23320 (cs) [Submitted on 26 Feb 2026] Title:ParamMem: Augmenting Language Agents with Parametric Reflective Memory Authors:Tianjun Yao, Yongqiang Chen, Yujia Zheng, Pan Li, Zhiqiang Shen, Kun Zhang View a PDF of the paper titled ParamMem: Augmenting Language Agents with Parametric Reflective Memory, by Tianjun Yao and 5 other authors View PDF HTML (experimental) Abstract:Self-reflection enables language agents to iteratively refine solutions, yet often produces repetitive outputs that limit reasoning performance. Recent studies have attempted to address this limitation through various approaches, among which increasing reflective diversity has shown promise. Our empirical analysis reveals a strong positive correlation between reflective diversity and task success, further motivating the need for diverse reflection signals. We introduce ParamMem, a parametric memory module that encodes cross-sample reflection patterns into model parameters, enabling diverse reflection generation through temperature-controlled sampling. Building on this module, we propose ParamAgent, a reflection-based agent framework that integrates parametric memory with episodic and cross-sample memory. Extensive experiments on code generation, mathematical reasoning, and multi-hop question answering demonstrate consistent improvements over state-of-the-art baselines. Further analysis reveals that ParamMem is sample-efficient, enables weak-to-strong transfer ...
