Saket Dingliwal

Finding Input Characterizations for Output Properties in ReLU Neural Networks

2020 · arXiv (Cornell University)

Deep Neural Networks (DNNs) have emerged as a powerful mechanism and are being increasingly deployed in real-world safety-critical domains. Despite the widespread success, their complex architecture makes proving any formal guarantees about them difficult. Identifying …

Think Clearly: Improving Reasoning via Redundant Token Pruning

2025 · arXiv (Cornell University)

Recent large language models have shown promising capabilities in long-form reasoning, following structured chains of thought before arriving at a final answer. However, we observe that these reasoning paths tend to include substantial redundancy; analyzing …