small language model

SolidGeo-Bench: A Pilot Study on Category-Level LLM Performance in Chinese High School Solid Geometry

Large language models (LLMs) are increasingly used in K-12 mathematics education, yet little is known about how their performance in solid geometry — a domain requiring spatial visualisation, cross-section reasoning and multi-step proof — varies across problem types. We present SolidGeo-Bench, a pilot set of 95 Chinese...

OSIRIS v4.5.2: a governed, local-first console for testing whether a small language model learns from conversation

v4.5.2 corrects how the core is scored. Until this release a layer of the core (the phase-conjugate corrector) ran during training in every block but was skipped whenever gradients were off -- in every held-out score, in early stopping and in generation -- so every held-out figure on record, including those below, desc...

EdgePulse: A Runtime Observability Framework for Quantized Large Language Models on Consumer Edge Devices

On-device AI execution lacks unified runtime observability across mobile platforms. Existing tools focus on isolated model profiling or single-engine metrics and give no synchronized view of memory, thermal state, battery draw, and inference latency during real workloads. We present EdgePulse, an open-source runtime ob...

Large language models as sources of diabetic retinopathy patient education: a comparative evaluation of safety, health information quality, and readability

Patients increasingly use large language models (LLMs) to seek medical information, but their performance in answering diabetic retinopathy (DR) questions remains insufficiently characterized. To evaluate the safety, accuracy, patient-centered communication, information quality, transparency, and readability...

Structured Pruning of Small Language Models: An Empirical Study on GQA-Aware Attention and MLP Compression with LoRA Recovery

Large language models have shown strong performance across many natural language tasks, but their size makes deployment difficult on hardware with limited memory and compute. Compression techniques aim to reduce this cost while preserving as much quality as possible. This study investigates structured pruning applied t...

Music therapy and music-based support for autistic children: a neurodiversity-informed systematic review and meta-analysis

Music-based interventions, including music therapy, are widely used to support autistic children, but reductions on autism-related scales do not fully represent communication, participation, quality of life, or autonomy. We retained clinically interpretable outcomes while using a neurodiversity-informed framework t...

Citation Fabrication and Accuracy in AI-Generated Psychiatric Research: A Comparative Evaluation of Four Contemporary Large Language Models

Objectives. LLMs are increasingly used to support academic writing, but their tendency to hallucinate citations raises concerns about the integrity of AI-assisted scholarly work. This study provides a timely snapshot of citation reliability in LLM-generated academic reviews by testing whether citation fabrication and a...

Removable and Insertable Mechanisms for Small Language Models

Can a small language model benefit from simple, hand-written tools for tasks such as copying a pattern or tracking brackets? We pair these tools with a standard transformer and test whether the model uses them by removing, swapping, or adding them. On structured synthetic tasks, explicit modules outperform tested gener...

A Two-Branch Hybrid Model for Imbalanced Lumbar Disc Grading and Zero-Shot Label Extension

An Overview of Knowledge-Based System: History, Advances and Challenges in the Era of Large Language Models

With 60 years of research, knowledge-based system has travelled a long journey from early small-scale rule-based programs to modern super-large-scale knowledge graphs. Several multithreading technical stages can be broadly categorized: early history of rule-based system, knowledge engineering, semantic network, semanti...

Spatial cognitive patterns of underground public spaces from a stimulus-organism-response perspective: evidence from LLM-empowered social media data mining

Current underground public space (UPS) renewal prioritizes engineering-oriented physical upgrades, neglecting users’ multidimensional cognitive experiences. Existing studies on underground space perceptions rely on small-sample surveys and fail to capture joint co-occurrence patterns among multiple stimuli. This study...

OSIRIS v4.5.3: a governed, local-first console for testing whether a small language model learns from conversation

v4.5.3 is a repair release. The exchange ledger's writer now chains each entry to the ledger file's last entry under a lock, so two consoles can no longer fork the hash chain; the chain check reports forks separately from damage (the author's ledger: 123 entries intact, seven forks from 2026-10-07, nothing rewritten)....