← Search

Manyi Zhang

2 accepted papers

2025

L-MTP: Leap Multi-Token Prediction Beyond Adjacent Context for Large Language Models

NeurIPS 2025poster

Large language models (LLMs) have achieved notable progress. Despite their success, next-token prediction (NTP), the dominant method for LLM training and inference, is constrained in both contextual coverage and inference efficiency due to its inherently sequential process. To overcome these challen…

Cited by 0SourcecodeScholar
2023

When Noisy Labels Meet Long Tail Dilemmas: A Representation Calibration Method

ICCV 2023oral

Real-world large-scale datasets are both noisily labeled and class-imbalanced. The issues seriously hurt the generalization of trained models. It is hence significant to address the simultaneous incorrect labeling and class-imbalance, i.e., the problem of learning with noisy labels on long-tailed da…

Cited by 25PDFcodeScholar