Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
Affiliation:
2
0
2
Pre-trained tabular foundation models unlock superior anomaly detection with minimal labeled data, outperforming domain-specific methods even with only eight examples.
PhySciBench reveals that top LLMs struggle with scientific reasoning, achieving only 33.5% accuracy, while DelveAgent demonstrates a promising 7.5% improvement in performance.