Category
2,967 papers
This work proposes a framework for locally masked SAE-based anomaly detection, supported by theoretical justifications, and demonstrates their ability to capture meaningful safety information while using only 1-2% of SAE neurons for computation.
FARSIGHT (Financial Agent Robustness and Security Investigation and Global Holistic Testing), a framework that performs scheme-level evaluation of financial LLM agents on two axes: robustness under market turbulence and security against three attack types: attacks on information sources, attacks on agents, and agent-as...
Depth-recurrent language models iteratively apply a small layer stack, decoupling per-token compute from distinct parameter count. To determine whether such a model genuinely utilizes its depth, both recurrence and layer-pruning literatures rely on a shared evaluation: truncating depth at inference time, plotting quali...
This work introduces QVAC Genesis III, a 191.43B-token, STEM-focused multi-domain synthetic corpus covering 19 domains across several difficulty levels and different educational styles, and introduces an LLM-as-a-parser evaluation protocol that extracts final answers from free-form outputs and tracks both accuracy and...
Reach audiences
Advertise in front of researchers, engineers, and readers.
ABSTRACT Local wisdom-based oral literature learning needs to be designed not only to preserve cultural knowledge but also to provide opportunities for students to develop digital literacy and scientific communication skills. This study aims to develop oral literature learning through digital storytelling media integr...
Generative artificial intelligence has entered medicine through a heterogeneous set of uses rather than through a single technological pathway. Large language models now draft clinical text, interpret reports, answer patient questions, support knowledge retrieval, structure health data, assist education, and participat...
Despite growing interest in LLM chatbots, little evidence exists on usage patterns in hospital systems. This paper presents a mixed-methods case study of “InternalGPT,” a HIPAA-compliant LLM chatbot made available to employees of the institution. Survey responses and 14 months of system utilization data were analyzed a...
A simple generative approach for creating synthetic 5′-UTR libraries based solely on the genomic sequence statistics of any desired organism and demonstrates that simple statistical language model approaches applied to genomic data can generate functional translational regulatory sequence libraries without detailed mec...
Public media communication, particularly through newspapers, plays a critical role in shaping public awareness, governmental response, and remediation efforts following large-scale crises, including nuclear power plant accidents. Because children are biologically more vulnerable to the effects of ionizing radiation, un...
The present study investigates the relationship between first language (L1) and second language (L2) oral fluency and their association with comprehensibility in lower-intermediate Japanese learners of English, examining which L2 fluency features are most closely associated with L1 English listeners' comprehensibility...
Few-shot prompting heavily improves query generation quality and enables effective use of smaller models, supporting the practical deployment of locally hosted LLMs for knowledge graph question answering.
StudentSim is a research system for evaluating and improving AI tutoring tools before they are tested with learners. It models how individuals may engage with tutors, giving feedback to improve tutor performance across subjects like chess, writing, and mathematics. The post Simulating student learning for better AI tutoring appeared first on Microsoft Research.
Earth AI
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
