Apps
Interactive companions to my papers.
EMNLP 2026 Industry Track
ForensicBench
A benchmark for agentic LLMs on journal-entry fraud detection: the task, the ledgers, the reference agent, the results, and an open leaderboard with automatic scoring.
IJCAI 2026 FINLLM
CI-FSFD Demo
An interactive explorer for the CI-FSFD benchmark of financial statement fraud detection: data pipeline, dataset samples, prompts, fine-tuning architecture and results.