Picture for Mengzhou Xia

Mengzhou Xia

BRIGHT: A Realistic and Challenging Benchmark for Reasoning-Intensive Retrieval

Add code
Jul 16, 2024
Viaarxiv icon

CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs

Add code
Jun 26, 2024
Figure 1 for CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs
Figure 2 for CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs
Figure 3 for CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs
Figure 4 for CharXiv: Charting Gaps in Realistic Chart Understanding in Multimodal LLMs
Viaarxiv icon

SimPO: Simple Preference Optimization with a Reference-Free Reward

Add code
May 23, 2024
Viaarxiv icon

Lory: Fully Differentiable Mixture-of-Experts for Autoregressive Language Model Pre-training

Add code
May 06, 2024
Viaarxiv icon

What's in Your "Safe" Data?: Identifying Benign Data that Breaks Safety

Add code
Apr 01, 2024
Viaarxiv icon

LESS: Selecting Influential Data for Targeted Instruction Tuning

Add code
Feb 20, 2024
Viaarxiv icon

Language Models as Science Tutors

Add code
Feb 16, 2024
Viaarxiv icon

Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications

Add code
Feb 07, 2024
Viaarxiv icon

Detecting Pretraining Data from Large Language Models

Add code
Nov 03, 2023
Viaarxiv icon

Catastrophic Jailbreak of Open-source LLMs via Exploiting Generation

Add code
Oct 10, 2023
Viaarxiv icon