Search papers, labs, and topics across Lattice.
The paper introduces ParamMem, a parametric memory module that encodes cross-sample reflection patterns into model parameters to promote diverse reflection generation via temperature-controlled sampling in language agents. They found a strong correlation between reflective diversity and task success, motivating the design of ParamMem to address the issue of repetitive outputs in self-reflection. Integrating ParamMem into a reflection-based agent framework (ParamAgent) leads to performance gains on code generation, mathematical reasoning, and multi-hop question answering tasks compared to SOTA baselines.
Language agents can achieve more diverse and effective self-reflection by encoding cross-sample reflection patterns directly into model parameters, leading to significant performance gains in reasoning tasks.
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 across model scales, and supports self-improvement without reliance on stronger external model, highlighting the potential of ParamMem as an effective component for enhancing language agents.
