Future Is Unevenly Distributed: Forecasting Ability of LLMs Depends on What We're Asking

Cornell Tech has secured a $7 million investment from NASA and Schmidt Sciences aimed at modernizing arXiv, a preprint repository for scientific papers. This funding will facilitate the migration of arXiv to cloud infrastructure, upgrade its outdated codebase, and develop new tools to enhance the discovery of relevant preprints for researchers.

A novel approach named PocketLLM has been introduced to address the challenges of compressing large language models (LLMs) for efficient storage and transmission on edge devices. This method utilizes meta-networks to project LLM weights into discrete latent vectors, achieving significant compression ratios, such as a 10x reduction for Llama 2-7B, while maintaining accuracy.

A recent analysis has been conducted on semi-supervised learning within hypergraphs, revealing that variational learning on random geometric hypergraphs can achieve asymptotic consistency. This study introduces Higher-Order Hypergraph Learning (HOHL), which utilizes Laplacians from skeleton graphs to enhance multiscale smoothness and converges to a higher-order Sobolev seminorm, demonstrating strong empirical performance on standard benchmarks.

A new framework called Task-aware Virtual View Exploration (TVVE) has been introduced to enhance robotic manipulation by integrating virtual view exploration with task-specific representation learning. This approach addresses limitations in existing vision-language-action models that rely on static viewpoints, improving 3D perception and reducing task interference.

A recent study highlights the limitations of evaluating machine unlearning (MU) by relying solely on a single training seed, revealing that results can vary significantly based on the random number seed used during model training. This finding emphasizes the need for more robust empirical comparisons in MU algorithms, particularly those that are deterministic in nature.

A recent study evaluated the mathematical reasoning capabilities of Large Language Models (LLMs) using the 2026 Korean College Scholastic Ability Test (CSAT) Mathematics section, ensuring a contamination-free evaluation environment. The research involved digitizing all 46 questions immediately after the exam's public release, allowing for a rigorous assessment of 24 state-of-the-art LLMs across various input modalities and languages.

PRISM-Bench has been introduced as a new benchmark for evaluating multimodal large language models (MLLMs) through puzzle-based visual tasks that assess both problem-solving capabilities and reasoning processes. This benchmark specifically requires models to identify errors in a step-by-step chain of thought, enhancing the evaluation of logical consistency and visual reasoning.

A recent position paper emphasizes the need for literary scholars to engage with research on large language model (LLM) interpretability, suggesting that the red team could serve as a platform for this ideological struggle. The paper argues that current interpretability standards are insufficient for evaluating LLMs.