In this paper, we delve into the computation graph of the language model to uncover the knowledge circuits that are instrumental in articulating specific knowledge.

Imitating humans' mental world knowledge model which provides global prior knowledge before the task and maintains local dynamic knowledge during the task, in this paper, we introduce parametric W orld K nowledge M odel ( WKM) to facilitate agent

Disentangled representation learning (DRL) aims to identify and decompose underlying factors behind observations, thus facilitating data perception and generation.

Lijun Zhang is the corresponding author.

Training offline RL models using visual inputs poses two significant challenges, i.e.,

We study online composite optimization under the Stochastically Extended Adversarial (SEA) model.

In this work, we present a dual-stage low-bit post-training quantization (PTQ) method for image super-resolution, namely 2DQuant, which achieves efficient and accurate SR under low-bit quantization.

Due to its powerful representational capabilities, Transformers have gradually become the mainstream model in the field of machine vision.