A technique referred to as ‘model pruning’ can be utilized to identify and eliminate neurons that repeatedly lead to skewed responses, according to researchers from Stanford Law School and the Stanford Institute for Human-Centered AI.