✨Nudging Hidden States: Training-Free Model Steering for Chain-of-Thought Reasoning in Large Audio-Language Models 📝 Summary: This paper introduces training-free inference-time model steering to enhance Chain-of-Thought reasoning in Large Audio-Language Models. It achieves accuracy gains up to 4.4% and shows cross-modal transfer, where text-derived steering vectors efficiently guide speech reasoning. This positions mode... 🔹 Publication Date: Published on Mar 15 🔹 Paper Links: • arXiv Page: https://arxiv.org/abs/2603.14636 • PDF: https://arxiv.org/pdf/2603.14636 ================================== For more data science resources: ✓ https://t.me/DataScienceT #AI #MachineLearning #LALMs #ChainOfThought #ModelSteering