1 citations · 2 across the 9 of their papers we have counts for
6 papers · 1 filter
AbsTopK: Rethinking Sparse Autoencoders For Bidirectional Features
Xudong Zhu, Mohammad Mahdi Khalili, Zhihui Zhu
Sparse autoencoders (SAEs) have emerged as powerful techniques for interpretability of large language models (LLMs), aiming to decompose hidden states into meaningful semantic feat…
From Emergence to Control: Probing and Modulating Self-Reflection in Language Models
Xudong Zhu, Jiachen Jiang, Mohammad Mahdi Khalili +1
Self-reflection -- the ability of a large language model (LLM) to revisit, evaluate, and revise its own reasoning -- has recently emerged as a powerful behavior enabled by reinforc…
Towards Understanding and Improving Refusal in Compressed Models via Mechanistic Interpretability
Vishnu Kabir Chhabra, Mohammad Mahdi Khalili
The rapid growth of large language models has spurred significant interest in model compression as a means to enhance their accessibility and practicality. While extensive research…
Post-processing for Fair Regression via Explainable SVD
Zhiqun Zuo, Ding Zhu, Mohammad Mahdi Khalili
This paper presents a post-processing algorithm for training fair neural network regression models that satisfy statistical parity, utilizing an explainable singular value decompos…
An Efficient Training Algorithm for Models with Block-wise Sparsity
Ding Zhu, Zhiqun Zuo, Mohammad Mahdi Khalili
Large-scale machine learning (ML) models are increasingly being used in critical domains like education, lending, recruitment, healthcare, criminal justice, etc. However, the train…
Neuroplasticity and Corruption in Model Mechanisms: A Case Study Of Indirect Object Identification
Vishnu Kabir Chhabra, Ding Zhu, Mohammad Mahdi Khalili
Previous research has shown that fine-tuning language models on general tasks enhance their underlying mechanisms. However, the impact of fine-tuning on poisoned data and the resul…