small language model

Internationalization of higher education in South Korea: trends, transformations, and emerging challenges

South Korea has pursued the internationalization of higher education through policies and institutional initiatives aimed at increasing international mobility, strengthening global competitiveness, and enhancing the international profile of its universities. This study examines how the internationalization of South Kor...

Replication package for "When Smaller Is Not Cheaper: The Verbosity Inversion in Large Language Model Cascades"

Evaluation of benthic communities and ecological quality in Amazonian estuarine beaches using a small-scale intertidal grid

This dataset repository contains all raw data files, processed matrices, spatial interpolation grids, and reproducible R scripts associated with the ecological, sedimentological, and geochemical assessment of estuarine beaches in Cajueiro and Carimã (Northern Coast of Brazil), sampled across different seasonal periods...

Deterministic Finite-Scale Geometry of RNA Language Model Embeddings Predicts siRNA Efficacy: Gϵ(r) Outperforms LID as a Functional Complexity Measure

Motivation. BLAST-based screening of small interfering RNAs (siRNAs) frequently produces false positives in the “grey zone”, where alignment statistics are marginally significant but functional affinity is low. Existing approaches based on local intrinsic dimensionality (LID) of language model embeddings are unstable o...

What Matters in RAG for Small Language Models? A Large-Scale Sensitivity Study

Guarantees, dispositions and answerable judges: An empirical basis for the behavioural certification of tool-using language agents

Buyers of tool-using language agents ask for evidence that an agent respects the authority it has been given, and what they receive is usually the vendor's own account. We ask what an independent behavioural certificate can claim about an agent the certifier cannot inspect, using a deployed certification suite as the i...

Semantic Fragmentation and Stochastic Assembly: A Protocol for Decentralized Language-Model Inference over Untrusted Volunteer Nodes

Peer-to-peer language-model inference has so far been pursued by splitting the model: transformer layers or tensors are distributed across machines, and intermediate activations traverse the public internet on every generated token. This places the design squarely against a bandwidth gap of roughly five orders of magni...

Genome‐wide association analysis of resistance to scald in an adapted multiparent winter malting barley population

ABSTRACT Scald, caused by the fungus Rhynchosporium graminicola Heinsen 1897, is a major foliar disease in winter malting barley ( Hordeum vulgare L). Resistance to scald in winter malting barley is controlled by major and minor resistance genes. We used a population of 377 lines derived from biparental crosses among f...

Language of Toxicity: An eXplainable Artificial Intelligence Approach

Toxicity prediction in small molecules represents a fundamental challenge in drug development and chemical safety assessment. Traditional approaches heavily rely on predefined molecular descriptors or fingerprints, potentially limiting the ability to capture complex and nonlinear structure–activity relationships. Here,...

EvoMO-SR: Multiobjective LLM-based Evolution of Symbolic Expressions with substructure guidance

EvoMO-SR is proposed, a novel LLM-driven SR framework in which the LLM generates equation skeletons, with their coefficients fitted separately by an external optimizer, which achieves the best accuracy in seven of the eight in-domain and out-of-domain settings for LSR-Synth.

FluxLite: Inference-Time Proposal Control for Discrete Diffusion Models

FluxLite is introduced, a lightweight, training-free proposal-control framework for discrete diffusion, identifying a tilted-path coverage factor that governs robustness to score error, together with finite-particle convergence for a fixed controlled Feynman-Kac recursion.

The Teacher Is a Direction, Not a Destination: Extrapolating RL-Induced Representation Residuals in On-Policy Distillation

Across four base/RL-teacher pairs spanning different scales, architectures, and pre-training lineages, RIDE approaches or exceeds the RL-trained teacher on every pair and is the only method whose mean does so, and it consistently outperforms output-space extrapolation, which degrades the student whenever the teacher is...