2 citations · 3 across the 11 of their papers we have counts for
12 papers
Token-Level Off-Policy Learning for Faithful Generation Under Distribution Shift
Zitong Huang, Gustavo Lucas Carvalho, Deqing Fu +1
We propose Token-Level Off-Policy Labeling (TOPL), an off-policy training paradigm that reframes post-training as a token-level correctness prediction task. Our key intuition is th…
Convergent Evolution: How Different Language Models Learn Similar Number Representations
Deqing Fu, Tianyi Zhou, Mikhail Belkin +2
Language models trained on natural text learn to represent numbers using periodic features with dominant periods at . In this paper, we identify a two-tiered hierarchy…
EPSVec: Efficient and Private Synthetic Data Generation via Dataset Vectors
Amin Banayeeanzade, Qingchuan Yang, Deqing Fu +6
High-quality data is essential for modern machine learning, yet many valuable corpora are sensitive and cannot be freely shared. Synthetic data offers a practical substitute for do…
Transformers Provably Learn Algorithmic Solutions for Graph Connectivity, But Only with the Right Data
Qilin Ye, Deqing Fu, Robin Jia +1
Transformers often fail to learn generalizable algorithms, instead relying on brittle heuristics. Using graph connectivity as a testbed, we explain this phenomenon both theoretical…
LLM Unlearning Without an Expert Curated Dataset
Xiaoyuan Zhu, Muru Zhang, Ollie Liu +2
Modern large language models often encode sensitive, harmful, or copyrighted knowledge, raising the need for post-hoc unlearning-the ability to remove specific domains of knowledge…
Zebra-CoT: A Dataset for Interleaved Vision Language Reasoning
Ang Li, Charles Wang, Deqing Fu +9
Humans often use visual aids, for example diagrams or sketches, when solving complex problems. Training multimodal models to do the same, known as Visual Chain of Thought (Visual C…
Textual Steering Vectors Can Improve Visual Understanding in Multimodal Large Language Models
Woody Haosheng Gan, Deqing Fu, Julian Asilis +5
Steering methods have emerged as effective and targeted tools for guiding large language models' (LLMs) behavior without modifying their parameters. Multimodal large language model…
FoNE: Precise Single-Token Number Embeddings via Fourier Features
Tianyi Zhou, Deqing Fu, Mahdi Soltanolkotabi +2
Large Language Models (LLMs) typically represent numbers using multiple tokens, which requires the model to aggregate these tokens to interpret numerical values. This fragmentation…
TLDR: Token-Level Detective Reward Model for Large Vision Language Models
Deqing Fu, Tong Xiao, Rui Wang +5
Although reward models have been successful in improving multimodal large language models, the reward models themselves remain brutal and contain minimal information. Notably, exis…
Pre-trained Large Language Models Use Fourier Features to Compute Addition
Tianyi Zhou, Deqing Fu, Vatsal Sharan +1
Pre-trained large language models (LLMs) exhibit impressive mathematical reasoning capabilities, yet how they compute basic arithmetic, such as addition, remains unclear. This pape…
IsoBench: Benchmarking Multimodal Foundation Models on Isomorphic Representations
Deqing Fu, Ruohao Guo, Ghazal Khalighinejad +5
Current foundation models exhibit impressive capabilities when prompted either with text only or with both image and text inputs. But do their capabilities change depending on the…
Transformers Learn to Achieve Second-Order Convergence Rates for In-Context Linear Regression
Deqing Fu, Tian-Qi Chen, Robin Jia +1
Transformers excel at in-context learning (ICL) -- learning from demonstrations without parameter updates -- but how they do so remains a mystery. Recent work suggests that Transfo…