3 papers
cs.AI2026
BenchBench-Protocol: Evaluating Real-World Wet-Lab Protocol Reasoning and Modification
Aditya Sivakumar, Ashu Singhal, Nicholas Larus-Stone +1
We introduce BenchBench-Protocol, a benchmark for large language models of 149 protocol-modification tasks recovered from modifications that scientists made to published protocols…
cs.LG2025
Enforcing Orderedness to Improve Feature Consistency
Sophie L. Wang, Alex Quach, Nithin Parsan +1
Sparse autoencoders (SAEs) have been widely used for interpretability of neural networks, but their learned features often vary across seeds and hyperparameter settings. We introdu…
q-bio.BM2025
Towards Interpretable Protein Structure Prediction with Sparse Autoencoders
Nithin Parsan, David J. Yang, John J. Yang
Protein language models have revolutionized structure prediction, but their nonlinear nature obscures how sequence representations inform structure prediction. While sparse autoenc…