2026-05
Paper
One paper accepted at TMLR
A new TMLR paper extends the lab's work on provable and actionable explanations.
View related record News & Lab Notes
The PAI lab record.
Paper updates, service, demos, reading notes, open positions, and project explainers for provable, actionable interpretability.
2026-04
Paper
Eight papers accepted at ICML 2026
Congratulations to all students and co-authors on a broad set of ICML 2026 results.
View related record2026-04
Paper
Three papers accepted at ACL 2026
The accepted work spans large model behavior, safety, and representation analysis.
View related record2026-02
Service
Lijie Hu serves as Action Editor of DMLR
PAI contributes to the research community through editorial and reviewing service.
View related record2026-01
Paper
Six papers accepted at ICLR 2026
The ICLR papers connect interpretability, protein language models, sparse autoencoders, and data influence.
View related record2025-11
2025-09
Paper
EAP-GP accepted at NeurIPS 2025
The work studies saturation effects in gradient-based automated circuit identification.
View related record2025-08
Lab
PAI begins at MBZUAI
Lijie Hu starts as a tenure-track Assistant Professor of Machine Learning at MBZUAI.
View related record