Picture for Chengxi Li

Chengxi Li

Zhejiang University

From Data-Centric to Sample-Centric: Enhancing LLM Reasoning via Progressive Optimization

Add code
Jul 09, 2025
Viaarxiv icon

Efficient and Adaptive Simultaneous Speech Translation with Fully Unidirectional Architecture

Add code
Apr 16, 2025
Viaarxiv icon

LLMs Can Achieve High-quality Simultaneous Machine Translation as Efficiently as Offline

Add code
Apr 13, 2025
Viaarxiv icon

The Curse of CoT: On the Limitations of Chain-of-Thought in In-Context Learning

Add code
Apr 07, 2025
Viaarxiv icon

Street Gaussians without 3D Object Tracker

Add code
Dec 07, 2024
Figure 1 for Street Gaussians without 3D Object Tracker
Figure 2 for Street Gaussians without 3D Object Tracker
Figure 3 for Street Gaussians without 3D Object Tracker
Figure 4 for Street Gaussians without 3D Object Tracker
Viaarxiv icon

GFreeDet: Exploiting Gaussian Splatting and Foundation Models for Model-free Unseen Object Detection in the BOP Challenge 2024

Add code
Dec 03, 2024
Viaarxiv icon

OnlySportsLM: Optimizing Sports-Domain Language Models with SOTA Performance under Billion Parameters

Add code
Aug 30, 2024
Figure 1 for OnlySportsLM: Optimizing Sports-Domain Language Models with SOTA Performance under Billion Parameters
Figure 2 for OnlySportsLM: Optimizing Sports-Domain Language Models with SOTA Performance under Billion Parameters
Figure 3 for OnlySportsLM: Optimizing Sports-Domain Language Models with SOTA Performance under Billion Parameters
Figure 4 for OnlySportsLM: Optimizing Sports-Domain Language Models with SOTA Performance under Billion Parameters
Viaarxiv icon

Step-level Value Preference Optimization for Mathematical Reasoning

Add code
Jun 16, 2024
Viaarxiv icon

AlphaMath Almost Zero: process Supervision without process

Add code
May 06, 2024
Figure 1 for AlphaMath Almost Zero: process Supervision without process
Figure 2 for AlphaMath Almost Zero: process Supervision without process
Figure 3 for AlphaMath Almost Zero: process Supervision without process
Figure 4 for AlphaMath Almost Zero: process Supervision without process
Viaarxiv icon

MARIO Eval: Evaluate Your Math LLM with your Math LLM--A mathematical dataset evaluation toolkit

Add code
Apr 22, 2024
Viaarxiv icon