From Projection to Prediction: Beyond Logits for Scalable Language Models

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 This funding will facilitate the migration of arXiv to cloud infrastructure, upgrade its outdated codebase, and develop new tools to enhance the discovery of relevant preprints for researchers.

Recent research has introduced the concept of representational stability in large language models (LLMs), focusing on how these models encode distinctions between true, false, and neither-true-nor-false content. The study assesses this stability by training a linear probe on LLM activations to differentiate true from not-true statements and measuring shifts in decision boundaries under label changes.

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 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.

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 new framework named Reading Comprehension Exercise Generation (RCEG) has been proposed to leverage large language models (LLMs) for automatically generating personalized English reading comprehension exercises. This framework utilizes fine-tuned LLMs to create content candidates, which are then evaluated by a discriminator to select the highest quality output, significantly enhancing the educational content generation process.

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 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.