COMPASS: Context-Modulated PID Attention Steering System for Hallucination Mitigation

Large language models (LLMs) have demonstrated remarkable abilities in natural language processing, image recognition, and multimodal tasks, positioning them as pivotal in the technological landscape. This article reviews the foundational technologies behind LLMs, such as prompt engineering and fine-tuning, while also exploring the historical evolution of the Science of Science (SciSci). It anticipates future applications of LLMs in scientometrics and discusses the potential of AI-driven models for scientific evaluation.

Large language models (LLMs) often generate confident yet incorrect outputs, necessitating reliable uncertainty quantification. This study introduces information-lift certificates that compare model probabilities to a skeleton baseline, utilizing sub-gamma PAC-Bayes bounds effective under heavy-tailed distributions. The method achieved 77.0% coverage at 2% risk across eight datasets, significantly outperforming entropy-based methods in blocking critical errors, making it practical for real-world applications.

NVIDIA's Learn-to-Steer is set to address a significant limitation in text-to-image diffusion models, which struggle with basic spatial reasoning. These models can create photorealistic images but often misplace objects in relation to one another, such as placing a dog to the left of a teddy bear instead of the right. This advancement aims to enhance the accuracy of generated images by improving spatial understanding.

Large language models (LLMs) exhibit label length bias, where labels of varying lengths are treated inconsistently despite normalization efforts. This paper introduces normalized contextual calibration (NCC), a method that normalizes predictions at the full-label level, effectively addressing this bias. NCC demonstrates statistically significant improvements across multiple datasets and models, achieving up to 10% gains in F1 scores. Additionally, it extends bias mitigation to tasks like multiple-choice question answering, showing reduced sensitivity to few-shot example selection.

The paper titled 'SymLoc: Symbolic Localization of Hallucination across HaluEval and TruthfulQA' addresses the issue of hallucination in large language models (LLMs). It highlights the challenges LLMs face when encountering symbolic triggers, such as modifiers and named entities, which can lead to hallucinations. The authors propose a novel symbolic localization framework that utilizes linguistic and semantic knowledge to trace the origins of these hallucinations within the model, aiming to improve understanding and management of this phenomenon.