EMNLP: Unsupervised Concept Vector Extraction for Bias Control in LLMs

LLM, Hannah Cyberey, Yangfeng Ji, activation steering, bias, fairness

Our paper on extracting concept vectors for LLMs was presented at the 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP):

Steering “gender” concept in QWEN-1.8B, evaluated on an example from Winogenerated fill-in- the-blank task. Baseline shows the original probabilities with no steering applied.

All Posts by Category or Tags.