DrugRAG: Enhancing Pharmacy LLM Performance Through A Novel Retrieval-Augmented Generation Pipeline

OpenAI has launched GPT-5.2, its latest AI model, which is designed to enhance productivity and has shown mixed results in tests compared to its predecessor, GPT-5.1. This development comes amid increasing competition from Google's Gemini 3, which has rapidly gained a significant user base.

The introduction of Time Puzzles marks a significant advancement in evaluating iterative temporal reasoning in large language models (LLMs). This task combines factual temporal anchors with cross-cultural calendar relations, generating puzzles that challenge LLMs' reasoning capabilities. Despite the simplicity of the dataset, models like GPT-5 achieved only 49.3% accuracy, highlighting the difficulty of the task.

A recent study has proposed enhancements to zero-shot recognition of Activities of Daily Living (ADLs) using Large Language Models (LLMs) by implementing event-based segmentation and a novel method for estimating prediction confidence. This approach aims to improve the accuracy of sensor-based recognition systems in smart homes, which are crucial for applications in healthcare and safety management.

A new framework called From Rows to Reasoning (FRTR) has been introduced to enhance the reasoning capabilities of Large Language Models (LLMs) when dealing with complex spreadsheets. This framework includes FRTR-Bench, a benchmark featuring 30 enterprise-grade Excel workbooks, which aims to improve the understanding of multimodal data by breaking down spreadsheets into granular components.

Recent research indicates that in adapter-based vision-language models, the alignment of image and text representations occurs from the very first layer, challenging the previous understanding that such alignment is only evident in later layers. This was demonstrated using a novel synthesis method inspired by DeepDream, which successfully generated images that reflect salient features of textual concepts from the initial layer.

A new benchmark called KidVis has been introduced to evaluate the visual perceptual capabilities of Multimodal Large Language Models (MLLMs), specifically assessing their performance against that of 6-7 year old children across six atomic visual capabilities. The results reveal a significant performance gap, with human children scoring an average of 95.32 compared to GPT-5's score of 67.33.