From Interpretability to Control: Insights from Six Years of the TrustNLP WorkshopPublished in arXiv (submitted to EACL 2027), 2026 Previous Next