This work led by @jialicheng123 moves machine unlearning toward a more mechanistic understanding of why some examples, features, or concepts are easier or harder to remove. Today at 2pm PTD:
Excited that our #ACL2026 paper with zihengchen, @_cagarwal, @amirieb on Measuring Unlearning Difficulty with MechInterp!View quoted post on X
Time: Sun July 5, 14:00-15:30, Poster Session B
Location: Grand Hall
Looking forward to connecting with everyone!
aclanthology.org/2026.findings-…