Search papers, labs, and topics across Lattice.
This paper introduces SALA, a Semantic-Aware Logical Alignment framework that enhances in-context learning for complex reasoning by learning task-specific reasoning operations instead of relying on fixed demonstrations. By embedding these operations in a continuous semantic space and employing dynamic time warping (DTW) for alignment, SALA enables flexible matching of reasoning sequences, addressing the limitations of traditional retrieval methods. Experimental results show that SALA significantly outperforms existing demonstration selection techniques across multiple reasoning benchmarks and large language models (LLMs).
SALA achieves superior reasoning alignment by learning flexible, task-specific operations, outperforming traditional methods that rely on rigid matching criteria.
Effective in-context learning (ICL) for complex reasoning relies on selecting the right demonstrations. Traditional retrieval methods based on surface similarity fail to capture the underlying problem-solving logic. Recent logic-based methods address this by matching predefined reasoning steps, but the rigid rules and exact-match criteria is improper to handle flexible or diverse reasoning processes. To address the problem, we propose SALA, a Semantic-Aware Logical Alignment framework. Instead of relying on a fixed inventory, SALA automatically learns task-specific reasoning operations. It then embeds these operations into a continuous semantic space and uses dynamic time warping (DTW) to align the reasoning sequences. This approach allows for soft, flexible matching of reasoning logic while remaining highly interpretable. Experiments across four reasoning benchmarks and three LLMs demonstrate that SALA outperforms existing demonstration selection methods. Further analysis confirms the roles of the operation induction and the logical semantic alignment.
