Latest Articles for Data Evaluation

A new benchmark sheds light on hallucination in vision language models.

This study highlights the importance of dataset granularity in improving image-text retrieval systems.

Introducing an efficient way to evaluate the quality of generated samples using latent density scores.

A new benchmark improves models' understanding of long videos and language.

HaloQuest addresses hallucination issues in vision-language models with a new dataset.

A new benchmark seeks to enhance evaluations of OIE systems for better performance insights.

A new benchmark to test visual-language models on minimal changes in images and captions.

This study highlights the need for LLMs to know when to abstain.

Proper scoring rules enhance the evaluation of probabilistic forecasts across various fields.

A framework for better estimating treatment effects in paired cluster-randomized experiments.

Using AI-generated relevance marks for efficient evaluation of information retrieval systems.

A new method improves evaluation accuracy in authorship verification by reducing topic leakage.

A new framework enhances evaluation of RAG systems in specialized domains.

New methods offer better evaluation of language understanding in models.

MicroSSIM enhances image quality assessment in microscopy for better scientific outcomes.

A new framework for assessing the performance of RAG systems.

ArabLegalEval assesses LLMs' performance in handling Arabic legal information.

New benchmark tackles relation hallucinations in multimodal large language models.

A novel approach to assess health-related answers generated by AI models.

Soda-Eval sets new standards for chatbot evaluation methods.

A new benchmark and dataset enhance evaluation of medical language models.

A new approach to assessing how citations support statements in generated text.

Researchers examine the reliability of metrics for language model safety.

A multi-domain benchmark assesses LLMs' code generation abilities across various fields.

A new system optimizes AI responses for legal fields, focusing on New York City's Local Law 144.

A study on the effectiveness of image matching methods in diverse scenarios.

Examining LVLMs' effectiveness in generating multilingual art explanations.

This study evaluates how well AI categorizes images compared to humans.

A fresh evaluation method for large language models using nested API calls.

OpenACE provides a fair benchmark for assessing audio codecs across various conditions.

Learn how to evaluate and compare images effectively.

VERA enhances the accuracy and relevance of language model responses.

RAGProbe automates the evaluation of RAG systems, improving their performance and reliability.

A new dataset enhances evaluation of language models in clinical trial accuracy.

A dataset helps AI systems learn better from distracting visuals.

A study on how models follow instructions during complex dialogues.

HealthQ evaluates AI's ability to ask questions in patient care.

Exploring methods to improve multimodal models in breaking down visual questions.

Introducing MemSim, a tool for assessing memory effectiveness in language model assistants.

Introducing a new model and benchmark for evaluating multi-audio tasks.