small language model

Beyond Knowledge: CHORDI as a Meta-Reasoning Framework for Responsible Reasoning and Action in Autonomous AI

Abstract Autonomous AI systems require more than knowledge and task optimization: they must also determine when a goal, value, safety consideration, or human interest creates a conflict that warrants reconsideration of the current evaluative frame. Building on CHORDI (Conflict to Harmonization Operator via Re-Dimension...

A discrete relational model built from local difference, rotation, and closure — k-Flip Theory (A–E)

Dear Researcher, I am sharing a working research archive titled k-Flip Theory, now available on Zenodo as a five-part set (A–E). The project began from a deliberately simple question: If two physical contributions cancel to zero, what exactly has disappeared? The working premise is that zero net value need not erase th...

Sobriété Numérique and the Frugal Tutor: Environmental Justice, Accessibility, and the Cognitive Floor of Edge-Deployed Language Tutoring

BMO is an offline, voice-to-voice French tutoring agent designed to run entirely on consumer CPU hardware, built to test how far AI language tutoring can move away from centralized cloud APIs toward local, sovereign computation. This paper asks a specific empirical question: how small can a language model be shrunk bef...

Sobriété Numérique and the Frugal Tutor: Environmental Justice, Accessibility, and the Cognitive Floor of Edge-Deployed Language Tutoring

BMO is an offline, voice-to-voice French tutoring agent designed to run entirely on consumer CPU hardware, built to test how far AI language tutoring can move away from centralized cloud APIs toward local, sovereign computation. This paper asks a specific empirical question: how small can a language model be shrunk bef...

Projective Shannon Sketches for Transformer KV Memory: Needle Rescue under Compressed Keys on GPT Model

Projective KV-Cache: Slashing LLM Memory Costs by 50%–75% Without Context Amnesia (Empirical Validation on GPT-2) The Executive Problem: Large Language Models (LLMs) processing long documents face an expensive bottleneck: the Key-Value (KV) Cache. Storing every token's attention memory in GPU VRAM creates immense hardw...

Projective Shannon Sketches for Transformer KV Memory: Needle Rescue under Compressed Keys on GPT Model

Projective KV-Cache: Slashing LLM Memory Costs by 50%–75% Without Context Amnesia (Empirical Validation on GPT-2) The Executive Problem: Large Language Models (LLMs) processing long documents face an expensive bottleneck: the Key-Value (KV) Cache. Storing every token's attention memory in GPU VRAM creates immense hardw...

Artificial intelligence in non-small cell lung cancer: transforming diagnosis, treatment, and prognostic evaluation

Non-small cell lung cancer (NSCLC) is characterized by profound biological heterogeneity, frequently leading to late-stage diagnosis, recurrence, or metastasis. Conventional clinical pathways heavily depend on subjective visual interpretations, invasive tissue biopsies, and experience-based medical models that constrai...

What Is the Best AI Voice Recorder for Lectures?

Whether you're attending university classes, online courses, seminars, or professional training sessions, having a reliable way to capture every detail is essential. Many students are now looking for the best AI voice recorder, a lecture recorder and note taker, or simply a good voice recorder that can do more than rec...

Bencmarking Generative AI for Debugging Support in Program Education

Abstract—Programming activities, such as automatic debugging and program repair, are increasingly being supported by Generative Artificial Intelligence (GenAI). However, a rigorous empirical evaluation of the efficacy of various large language models (LLMs) across several debugging phases is necessary. Using a benchmar...

Fine-tuning small reasoning models for quantum field theory

Abstract Despite the growing application of large language models (LLMs) to theoretical physics, there has been little academic exploration of how domain-specific physics reasoning ability develops during training. To investigate this, we perform the first academic comparison of reinforcement learning and supervised fi...

Can a Small Model Learn to Look Before It Leaps? Dynamic Learning and Proactive Correction for Hallucination Detection

Solving the Needle-in-a-Haystack Problem in Mammography Vision-Language Model with Differentiable Subset Sampling

TopKSigLIP outperforms existing open-source mammography and general medical VLMs on both internal and external benchmarks on density assessment, BI-RADS classification, finding subtyping, and cancer prediction under zero-shot evaluation.