Latest Articles for Language Processing

LoRETTA improves fine-tuning efficiency for large language models with fewer parameters.

New methods to enhance continuous learning in language models while retaining past knowledge.

Language models excel at text but lack sensory understanding.

Study reveals language models prioritize relevance over evidence quality.

A new method helps robots follow complex commands more effectively.

This article examines how language models can adopt ideological biases from training data.

This article discusses a method to enhance language models using structured instructions.

Archer introduces complex reasoning to enhance text-to-SQL tasks in diverse languages.

Examining the combination of SFMs and LLMs for improved speech translation.

This study evaluates models for tracking shifts in word meanings across languages.

Examining the limitations of LLMs in understanding and retaining temporal information.

A new approach improves efficiency in multilingual ASR models by integrating adaptive masking techniques.

A new method improves alignment of LLMs with minimal human feedback.

Investigating deepfake audio to enhance transcription models for less common languages.

Exploring how tensor networks can enhance language modeling through Motzkin spin chains.

Study shows LLMs excel at answering from choices, revealing unexpected reasoning skills.

Exploring how word order influences language processing and communication.

Examining how new words affect language model performance.

SiLLM enhances real-time translation by integrating two distinct models.

Examining the sample sizes needed for specialized models to surpass general ones.

This article examines how restart-incremental models improve language understanding amidst local ambiguities.

Exploring in-context learning and its implications for multilingual AI performance.

Research on blending different communication styles in AI text generation.

This study examines the efficacy of multilingual models in following instructions across European languages.

A study on the role of Degenerate Knowledge Neurons in improving language model performance.

Investigating how tokenization methods affect arithmetic tasks in language models.

This article explores how language models can aid in writing academic meta-reviews.

A new framework enhances hate speech detection by generating realistic test cases.

An adaptive agent improves teamwork in Codenames using multiple language models.

A new method improves how AI models express their confidence in answers.

This article examines the dangers of harmful fine-tuning in language models.

A new approach using backtranslation aims to protect language models from harmful prompts.

A method for enhancing response quality in language models using feedback.

Study reveals challenges and progress in chatbot memory during lengthy dialogues.

Study evaluates LLMs' ability to create culturally relevant question-answer data.

This article examines the reliability of political views in large language models.

A new benchmark for assessing Korean conversational abilities of language models.

Discover why tokenization is key for computers to understand human language.

This study examines gender bias in large language models across multiple languages.

New methods aim to better evaluate reasoning skills in AI language models.