small language model

One Door, Three Modes: Designing a Multi-Channel Appointment Booking System Driven by Conversational Agents

Small service businesses such as clinics receive booking requests through WhatsApp, Instagram and Messenger messages, phone calls and web forms, and many have no calendar system at all. We describe the design of ABS, an appointment booking system in which conversational agents take bookings across these channels for ma...

Building an industry model framework with agentic AI and small language models (SLMs)

A Contamination Flag Is Not an Inflation Estimate: What Benchmark Contamination Detectors Certify, What Decontamination Is Meant to Buy, and Why the Clean-Subset Comparison That Joins Them Is Biased From Both Sides

(c) 2026 Pranay Mahendrakar. Licensed under CC BY 4.0. Almost every large language model report now includes a contamination analysis: benchmark items that overlap the training corpus are flagged, the model is rescored on the unflagged remainder, and a small gap is read as evidence that the headline numbers stand. Two...

EukaUTR: a foundation model unifying functional modelling and design of eukaryotic 3′ UTRs

Early Classification of Information Cascades with Class-Conditional Small Language Models

Early detection classifies an information cascade from its first nodes, but the labeled cascades available for training are complete. On four datasets, discriminative models trained on complete cascades are a median of 18 AUC points below their prefix-trained versions at five nodes. We write the structure and timing of...

Side channels in LLM secret defense: a pre-registered replication of the Paradox of Active Defense across five target models

Large language models deployed with a secret in their system prompt defend it only by following an instruction. A pilot study (88 tracked adversarial assessments against ten models and seven prompt designs, March to April 2026) found that the defense leaves traces while it holds: the model recites the secret in its rea...

goalbench: a reproducible benchmark of choice-style deciders for customer-goal classification in Brazilian Portuguese

goalbench measures small "decider" models that answer typed choice questions (Laya, Strands Decider) on customer-goal classification in Brazilian Portuguese, inside a two-stage pipeline: an embedding shortlist of 10 goals followed by the decider, with a calibrated combination fitted by cross-validation. Runs log every...

Data from: 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...

OSIRIS v4.5.0: a governed, local-first console for testing whether a small language model learns from conversation

v4.5.0 adds Google Cloud integration: Gemini through Vertex AI in the user's own project with their gcloud login, keys read from Google Secret Manager instead of a plaintext file, and CI on Google Cloud Build; it also includes NCLM-1 workflow and verify-loop changes. See RELEASE_NOTES_v4.5.0.md. OSIRIS is a local-first...

EL-RAKHAWI BEACON THE SYMBIOTIC CONVERGENCE SYSTEM FOR HUMAN SURVIVAL AND DIGNITY A COMPREHENSIVE TECHNICAL AND PHILOSOPHICAL FOUNDATION DOCUMENT FINAL HARDENED OPEN HARDWARE EDITION WITH REALISTIC ENGINEERING SPECIFICATIONS

EL-RAKHAWI BEACON: EXECUTIVE SUMMARY The El-Rakhawi Beacon is a symbiotic convergence system integrating four life-sustaining functions into a single solar-powered unit: atmospheric water generation (8-40 L/day via MOF sorbents), hybrid energy storage (LFP battery 5kWh), edge AI with mesh networking, and environmental-...

Testing six LLM-generated sleep-EEG predictions on open polysomnography: a case study with a pre-frozen protocol and negative results

Language models can write research hypotheses much faster than anyone can test them. Earlier studies have judged such hypotheses by expert ratings, benchmarks with known answers or laboratory experiments. We report a small, fully documented case from a different angle: what happens to a handful of model-generated predi...

Basic Concepts and Models for Interpreter and Translator Training

In the 1980s, Daniel Gile thought of the potential benefit for translation and interpreting students of simple conceptual building blocks that could help them make sense of their classroom experience and of their instructors’ advice, especially as regards recurring challenges. The first edition of Basic Concepts and Mo...