small language model

Compressing Streaming Neural Audio Encoders via Latent-Space Distillation

This work studies how to compress a tokenizer by distillation, taking as the supervision target neither the discrete token nor the output distribution but the pre-quantizer latent the model actually consumes - the last representation the two token interfaces share.

Learning curves and training in robotic organ transplantation: bibliometric mapping with kidney transplantation as the principal clinical model

Bibliometric mapping shows that the learning-explicit robotic transplantation literature is strongly kidney-dominant, with evidence from other organs representing smaller and clinically distinct extensions, and proposes a multidomain interpretation in which technical efficiency, transplant-specific process control, cli...

RAN-Doc: Retrieval-Augmented Small Language Models for Edge-Deployable Fault Diagnosis in AI-Native O-RAN

Domain-configured retrieval-grounded diagnosis from telemetry, alerts, and logs.

NGramLang: A Domain-Specific Language for Document Retrieval Using Character N-Grams and Okapi BM25 Ranking

NGramLang is a small domain-specific language and interpreter designed to make the mechanics of document retrieval explicit and executable. Through ten declarative commands, a user loads text documents, aggregates them into a corpus, and ranks them against a free-text query. The language front end is implemented with J...

The Language of Trust: The Roles of Service Discourse Markers in Shaping Guest Satisfaction and Perceived Professionalism in Tourism and Hospitality Encounters

The research reviews the literature on high-frequency Service Discourse Markers (SDMs) in the tourism and hospitality industries, including welcoming greetings, expressions of gratitude, apologies, small talk, hedges, and alignments that affect customer satisfaction and perceptions of professionalism. By regarding unde...

RAN-Doc: Retrieval-Augmented Small Language Models for Edge-Deployable Fault Diagnosis in AI-Native O-RAN

Domain-configured retrieval-grounded diagnosis from telemetry, alerts, and logs.

Pressure-Induced Directional Asymmetry in LLM Procedural Gating: DF-Core v1.3 Benchmark and Replication Package

Summary Benchmarks for policy-adherent language agents typically split tasks into those where the agent must refuse and those where it must act, and report accuracy on each. Such designs measure the joint effect of task difficulty and of whatever conversational dynamics the benchmark contains, without separating them....

RosettaMath: Reading, Running and Proving Mathematics from LATEX

The translation of theoretical mathematical models into executable code remains a persistentbottleneck in computational science. While LATEX serves as the standard for sharing algebraicequations and algorithms, its representation is fundamentally disconnected from the explicit logicrequired by programming languages lik...

The Tailor's Reading

A feature of a sparse-autoencoder dictionary is published with one natural-language label, written by one language model from the passages that most excite the feature in one corpus. That label is a reading, and its provenance, the explainer, the explainer's system prompt, the corpus, and the model through which the co...

Reliability Benchmark for Small Language model

Algebra Before Bit: A Computer Scientist's Account of a Finite Record Universe

Imagine inspecting a quantum experiment through a programmer's console: what must be stored, what changes when a detector is read, and what can its records tell us? This is a representation, not a hypothesis: nothing here says the universe is a simulation, and nothing here depends on it being one. We examine a proposed...

Reasoning Budget Sweep: truncation and answer-production measurements for 12 small reasoning LLMs

A per-response measurement dataset covering 12 open reasoning language models (0.6B to 7B) on AIME 2024, AIME 2025 and MATH-500 across six generation token budgets (512 to 16384) and three seeds. 11,118 scored responses. Each row records the token count, whether the response was truncated at the budget, whether an answ...