Our experts
AIM Faculty & Staff
Sarah Wiegreffe
Sarah Wiegreffe
Assistant Professor, Computer Science
Latest Papers
Are Latent Reasoning Models Easily Interpretable?
|
Third Conference on Language Modeling
Author(s): Dilgren, Connor, Wiegreffe, Sarah
UMD Author(s): Sarah Wiegreffe
A Closer Look at Failure Modes in Temporal Understanding of Large Audio-Language Models
|
Interspeech
Author(s): Kulkarni, Apoorva, Jayakumar, Kaousheik, Ghosh, Sreyan, et. al
UMD Author(s): Sarah Wiegreffe
Position: Interpretability Can Be Actionable
Author(s): Orgad, Hadas, Barez, Fazl, Haklay, Tal, et. al
UMD Author(s): Sarah Wiegreffe
Predictable Steering: Leveraging Geometric Proxies for Layer Selection and Per-Instance Success Prediction
|
Mechanistic Interpretability Workshop at ICML
Author(s): Hundia, Nishkal, Agrawal, Swastik, Goyal, Navita, et. al
UMD Author(s): Sarah Wiegreffe
Arguments that Alter Minds: LLM Rationales Sway Human (and LLM) Notions of Plausibility
|
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
UMD Author(s): Rachel Rudinger, Sarah Wiegreffe
Localizing Anchoring Pathways in Language Models
Author(s): Owusu, Hillary N., Wiegreffe, Sarah, Feldman, Naomi H.
UMD Author(s): Sarah Wiegreffe
Quantifying the Gap between Understanding and Generation within Unified Multimodal Models
|
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (Findings)
Author(s): Wang, Chenlong, Chen, Yuhang, Hu, Zhihan, et. al
UMD Author(s): Sarah Wiegreffe
Trait-space Monitoring for Emergent Misalignment During Supervised Finetuning
Author(s): Nghiem, Huy, Ho, Sy-Tuyen, Wiegreffe, Sarah, et. al
UMD Author(s): Sarah Wiegreffe
What Drives Representation Steering? A Mechanistic Case Study on Steering Refusal
Author(s): Cheng, Stephen, Wiegreffe, Sarah, Manocha, Dinesh
UMD Author(s): Sarah Wiegreffe