9 citations · 10 across the 3 of their papers we have counts for
3 papers · 1 filter
Dense SAE Latents Are Features, Not Bugs
Xiaoqing Sun, Alessandro Stolfo, Joshua Engels +4
Sparse autoencoders (SAEs) are designed to extract interpretable features from language models by enforcing a sparsity constraint. Ideally, training an SAE would yield latents that…
Efficient Annotator Reliability Assessment and Sample Weighting for Knowledge-Based Misinformation Detection on Social Media
Owen Cook, Charlie Grimshaw, Ben Wu +6
Misinformation spreads rapidly on social media, confusing the truth and targeting potentially vulnerable people. To effectively mitigate the negative impact of misinformation, it m…
Confidence Regulation Neurons in Language Models
Alessandro Stolfo, Ben Wu, Wes Gurnee +4
Despite their widespread use, the mechanisms by which large language models (LLMs) represent and regulate uncertainty in next-token predictions remain largely unexplored. This stud…