作者位次 #4/4ACL 2026Poster· Natural Language Processing
Revitalizing Black-Box Interpretability: Actionable Interpretability for LLM s via Proxy Models
Junhao Liu · Haonan Yu · Zhenyu Yan · Xin Zhang
摘要
Post-hoc explanations provide transparency and are essential for guiding model optimization, such as prompt engineering and data sanitation. However, applying model-agnostic techniques to Large Language Models (LLMs) is hindered by prohibitive computational costs, rendering these tools dormant fo...