1 citations · 3 across the 10 of their papers we have counts for
13 papers
Not All LLM Reasoning is Visible in the Chain-of-Thought
Vatsal Baherwani, Tom Goldstein, Ashwinee Panda
A key question for AI safety is whether a language model expresses all of its reasoning in its output tokens. We demonstrate a concrete failure mode where frontier models exhibit i…
Speculating Experts Accelerates Inference for Mixture-of-Experts
Vivan Madan, Prajwal Singhania, Abhinav Bhatele +2
Mixture-of-Experts (MoE) models have gained popularity as a means of scaling the capacity of large language models (LLMs) while maintaining sparse activations and reduced per-token…
From Refusal Tokens to Refusal Control: Discovering and Steering Category-Specific Refusal Directions
Rishab Alagharu, Ishneet Sukhvinder Singh, Shaibi Shamsudeen +2
Language models are commonly fine-tuned for safety alignment to refuse harmful prompts. One approach fine-tunes them to generate categorical refusal tokens that distinguish differe…
Multi-Token Prediction via Self-Distillation
John Kirchenbauer, Abhimanyu Hans, Brian Bartoldson +3
Existing techniques for accelerating language model inference, such as speculative decoding, require training auxiliary speculator models and building and deploying complex inferen…
FineGRAIN: Evaluating Failure Modes of Text-to-Image Models with Vision Language Model Judges
Kevin David Hayes, Micah Goldblum, Vikash Sehwag +3
Text-to-image (T2I) models are capable of generating visually impressive images, yet they often fail to accurately capture specific attributes in user prompts, such as the correct…
DynaGuard: A Dynamic Guardian Model With User-Defined Policies
Monte Hoover, Vatsal Baherwani, Neel Jain +7
Guardian models play a crucial role in ensuring the safety and ethical behavior of user-facing AI applications by enforcing guardrails and detecting harmful content. While standard…
Analysis of Attention in Video Diffusion Transformers
Yuxin Wen, Jim Wu, Ajay Jain +2
We conduct an in-depth analysis of attention in video diffusion transformers (VDiTs) and report a number of novel findings. We identify three key properties of attention in VDiTs:…
Dense Backpropagation Improves Training for Sparse Mixture-of-Experts
Ashwinee Panda, Vatsal Baherwani, Zain Sarwar +4
Mixture of Experts (MoE) pretraining is more scalable than dense Transformer pretraining, because MoEs learn to route inputs to a sparse set of their feedforward parameters. Howeve…
LoRI: Reducing Cross-Task Interference in Multi-Task Low-Rank Adaptation
Juzheng Zhang, Jiacheng You, Ashwinee Panda +1
Low-Rank Adaptation (LoRA) has emerged as a popular parameter-efficient fine-tuning (PEFT) method for Large Language Models (LLMs), yet it still incurs notable overhead and suffers…
Identifying and Evaluating Inactive Heads in Pretrained LLMs
Pedro Sandoval-Segura, Xijun Wang, Ashwinee Panda +4
Attention is foundational to large language models (LLMs), enabling different heads to have diverse focus on relevant input tokens. However, learned behaviors like attention sinks,…
Continual Pre-training of MoEs: How robust is your router?
Benjamin Thérien, Charles-Étienne Joseph, Zain Sarwar +7
Sparsely-activated Mixture of Experts (MoE) transformers are promising architectures for foundation models. Compared to dense transformers that require the same amount of floating-…
Gemstones: A Model Suite for Multi-Faceted Scaling Laws
Sean McLeish, John Kirchenbauer, David Yu Miller +5
Scaling laws are typically fit using a family of models with a narrow range of frozen hyperparameter choices. In this work we study scaling laws using multiple architectural shapes…
Refusal Tokens: A Simple Way to Calibrate Refusals in Large Language Models
Neel Jain, Aditya Shrivastava, Chenyang Zhu +6
A key component of building safe and reliable language models is enabling the models to appropriately refuse to follow certain instructions or answer certain questions. We may want…