4 articles tagged with "Reasoning"
The newly introduced VibeThinker model, with 3 billion parameters, demonstrates enhanced reasoning capabilities compared to Opus 4.5, marking a significant step in AI advancements.
The VAMPS benchmark sheds light on the performance of multimodal large language models in mathematical problem solving, revealing both capabilities and challenges.
A recent study highlights the challenges faced by large language models in performing structured logical reasoning, indicating key areas for improvement.
A recent study presents PathCal, a calibration method designed to improve the efficiency of Large Reasoning Language Models in complex reasoning tasks.