让提示词和模型参数协同优化，提升对话效率与准确性
Words & Weights: Streamlining Multi-Turn Interactions via Co-Adaptation
- 提出词与权重联合优化框架，同时修正意图模糊与能力不足
- 在MATH数据集上性能提升30%，交互轮次减少40%
- 适合需要高效动态调整的对话系统开发者
推理时策略自适应（T2PAM）对 align 大语言模型与动态用户需求至关重要。现有方法多将测试时适配视为单一维度问题，仅优化提示或仅调整权重，忽视了交互失败源于语义模糊与能力不足的耦合。我们提出ROSA2，将交互重构为词与权重的联合优化问题。通过数学分解误差信号，利用文本梯度修正意图模糊，通过参数更新弥补能力缺口。理论上证明该协同机制可严格减少收敛所需参数变化量。实验表明，ROSA2在MATH数据集上性能优于最先进基线30%，交互轮次减少40%，验证了优化上下文能释放参数更新的真实潜力。
原文摘要 · Abstract (English)
Test-time policy adaptation for multi-turn interactions (T2PAM) is essential for aligning Large Language Models (LLMs) with dynamic user needs during inference time. However, existing paradigms commonly treat test-time adaptation as a single-axis problem, either purely refining instructions (Prompt Engineering) or only adjusting weights (Test-Time Training), ignoring that interaction failures stem from a coupled mix of ambiguity and incapacity. We argue that these two optimization paths are not merely additive but synergistic: semantic clarity acts as a pre-conditioner for effective parameter updates. To this end, we propose ROSA2, a framework that reformulates interaction as a joint optimization problem over the heterogeneous space of Words and Weights. By mathematically decomposing the error signal, ROSA2 utilizes textual gradients to rectify intent ambiguity and parameter updates to bridge capability gaps. Theoretically, we prove that this co-adaptation strictly reduces the required parameter shift for convergence. Empirically, ROSA2 outperforms state-of-the-art baselines by 30% on MATH while reducing interaction turns by 40%, demonstrating that refining the context unlocks the true potential of parameter updates.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。