Latest Articles for Language Models

Examining vulnerabilities in watermarking methods against paraphrasing attacks.

Assessing language models' understanding of proverbs in low-resource languages.

Investigating how wealth influences language models in travel narratives.

Scar enhances language models by reducing toxic language in text generation.

Research shows variation in speech improves language model training.

Explore the impact of question styles on AI model performance.

A new method to develop guardrails for large language models without real-world data.

A new method enhances the safety of code generated by language models.

SpecTool brings clarity to LLM errors in using tools.

A study reveals how prompt injection can compromise language models.

This study examines how well LLMs assess creativity in the Alternative Uses Test.

PEFT methods enhance language models while safeguarding private data.

A study on how well language models connect facts without shortcuts.

A new method for language models to enhance their responses through self-generated critiques.

How low-bit quantization affects large language models during training.

A new method automates news classification, saving time and resources for organizations.

Evaluating if language models can understand spatial relationships effectively.

Discover how to improve large language models in handling symmetric tasks.

Evaluating language models' abilities in synthetic data creation using AgoraBench.

How language models improve their understanding of grammar and sentence structures.

Exploring how transformers can express uncertainty to improve AI reliability.

Large language models excel in some areas but struggle with general tasks.

Discover how activation sparsity boosts AI efficiency and speed.

Explore the connections between language models and physical phenomena in an engaging way.

Researchers are improving AI's ability to tackle complex questions with AutoReason.

Researchers tackle biases in language models for Filipino, enhancing cultural relevance.

This article examines the complex role of English in multilingual evaluations.

Learn how Sloth is changing predictions for language model performance.

BatchTopK sparse autoencoders improve language processing through smart data selection.

Addressing challenges in Multi-Hop Question Answering for better AI responses.

Combining language models enhances instruction-following in chip design tasks.

Smaller language models show surprising advantages in evolving instructions over larger ones.

Explore how memory structures enhance communication in large language models.

A tool that clarifies peer review comments for authors.

A new benchmark to test LLM reasoning across cultural backgrounds.

Learn how hints can improve AI reasoning and problem-solving skills.

Optimize your prompts with iPrOp for better AI responses.

Learn how external knowledge improves language model accuracy and reliability.

ChainRank improves text reranking, keeping models sharp and relevant.

LLMs can create tailored false content, increasing risks of deception.