163 citations · 165 across the 12 of their papers we have counts for
12 papers
Retrieval is Enough: Training-Free Interpretability with a Tool-Using Agent
Sriram Balasubramanian, Soheil Feizi
Interpretability methods for neural network activations span a wide cost spectrum, from cheap, training-free techniques (such as linear probes, PCA, SVD) to more expensive training…
Hop, Skip, and Overthink: Diagnosing Why Reasoning Models Fumble during Multi-Hop Analysis
Anushka Yadav, Isha Nalawade, Srujana Pillarichety +7
The emergence of reasoning models and their integration into practical AI chat bots has led to breakthroughs in solving advanced math, deep search, and extractive question answerin…
A Closer Look at Bias and Chain-of-Thought Faithfulness of Large (Vision) Language Models
Sriram Balasubramanian, Samyadeep Basu, Soheil Feizi
Chain-of-thought (CoT) reasoning enhances performance of large language models, but questions remain about whether these reasoning traces faithfully reflect the internal processes…
Tool Preferences in Agentic LLMs are Unreliable
Kazem Faghih, Wenxiao Wang, Yize Cheng +5
Large language models (LLMs) can now access a wide range of external tools, thanks to the Model Context Protocol (MCP). This greatly expands their abilities as various agents. Howe…
SpurLens: Automatic Detection of Spurious Cues in Multimodal LLMs
Parsa Hosseini, Sumit Nawathe, Mazda Moayeri +2
Unimodal vision models are known to rely on spurious correlations, but it remains unclear to what extent Multimodal Large Language Models (MLLMs) exhibit similar biases despite lan…
A Survey on Mechanistic Interpretability for Multi-Modal Foundation Models
Zihao Lin, Samyadeep Basu, Mohammad Beigi +18
The rise of foundation models has transformed machine learning research, prompting efforts to uncover their inner workings and develop more efficient and reliable applications for…
Decomposing and Interpreting Image Representations via Text in ViTs Beyond CLIP
Sriram Balasubramanian, Samyadeep Basu, Soheil Feizi
Recent work has explored how individual components of the CLIP-ViT model contribute to the final representation by leveraging the shared image-text representation space of CLIP. Th…
Rethinking Artistic Copyright Infringements in the Era of Text-to-Image Generative Models
Mazda Moayeri, Samyadeep Basu, Sriram Balasubramanian +4
Recent text-to-image generative models such as Stable Diffusion are extremely adept at mimicking and generating copyrighted content, raising concerns amongst artists that their uni…
Exploring Geometry of Blind Spots in Vision Models
Sriram Balasubramanian, Gaurang Sriramanan, Vinu Sankar Sadasivan +1
Despite the remarkable success of deep neural networks in a myriad of settings, several works have demonstrated their overwhelming sensitivity to near-imperceptible perturbations,…
C*-extreme entanglement breaking maps on operator systems
Sriram Balasubramanian, Neha Hotwani
Let denote the set of all unital entanglement breaking (UEB) linear maps defined on an operator system and, mapping into . As it turns ou…
Can AI-Generated Text be Reliably Detected?
Vinu Sankar Sadasivan, Aounon Kumar, Sriram Balasubramanian +2
Large Language Models (LLMs) perform impressively well in various applications. However, the potential for misuse of these models in activities such as plagiarism, generating fake…
Towards Improved Input Masking for Convolutional Neural Networks
Sriram Balasubramanian, Soheil Feizi
The ability to remove features from the input of machine learning models is very important to understand and interpret model predictions. However, this is non-trivial for vision mo…