Latest Articles for Language Models

Introducing RePrompt for better language model performance through optimized prompts.

A new benchmark evaluates how language models handle text changes.

User traits influence the responses of language models and their safety.

A toolkit for assessing performance of retrieval-augmented models in specific domains.

This study reveals how language models change behavior during training.

This article examines ways to improve planning abilities in large language models.

DetectBench evaluates LLMs on their ability to detect hidden evidence in reasoning tasks.

Examining how neuron activation enhances arithmetic reasoning in large language models.

A new model generates Czech poetry with improved rhyme and rhythm.

A new benchmark evaluates reasoning skills in language models.

A study on how language models generate persuasive rationales for argument evaluation.

This study assesses the honesty of LLMs in three key areas.

This article explores how adversaries impact teamwork among language models.

A comprehensive study on language models’ performance across 10 Indic languages.

A new method improves code repair for underused programming languages.

Exploring how attention sinks impact language model performance and introducing a calibration technique.

RankAdaptor optimizes fine-tuning for pruned AI models, enhancing performance efficiently.

A study on PlagBench and its role in detecting plagiarism in LLM outputs.

New dataset assesses LLMs' ability for complex logical reasoning tasks.

This research investigates how reasoning skills transfer across languages in language models.

This article discusses how AI models learn from mistakes through self-correction.

This study evaluates how well LLMs reason about cardinal directions.

This study assesses how well LLMs handle decision-making in a game setting.

Study reveals how user traits affect LLM responses and accuracy.

CharED combines language models for improved performance without shared vocabularies.

RAGBench introduces a comprehensive dataset for evaluating Retrieval-Augmented Generation systems.

Exploring fairness issues in AI language models and their implications.

Introducing a tool to enhance safety in language model interactions.

This article explores the detection of errors in tools used by language models.

This article analyzes repetitive structures in text generated by language models.

A new benchmark assesses how well language models follow multiple instructions in sequence.

MalAlgoQA dataset evaluates reasoning of Large Language Models in counterfactual scenarios.

MathCAMPS offers a fresh way to assess mathematical reasoning in language models.

This work focuses on better number representation using digit embeddings for improved predictions.

Exploring LLMs' effectiveness in decision-making through Dueling Bandits scenarios.

A new benchmark for assessing large language models in hypothesis testing.

CRAB enhances testing for language models in real-world environments.

Fine-tuning large language models directly on smartphones while protecting user data.

An overview of mechanistic interpretability in transformer-based language models.

Exploring how reframing shifts opinions through community discussions.