large language models

Artifact for T-REX: Teaching Large Language Models to Reason with Verbalized Execution Semantics

Artifact accompanying the paper T-REX: Teaching Large Language Models to Reason with Verbalized Execution Semantics. It contains the source code, training and evaluation scripts, configurations, and documentation required to reproduce the reported experiments.

Summarization Bias: The Directional Collapse of Objective Projection into Told-Mode Labels in Large Language Models — A Conceptual Framework and Registered Test Protocol

This paper introduces and operationalizes summarization bias: a proposed systematic tendency of large language models (LLMs) to represent narrative meaning as an abstract summary la-bel rather than as the reconstructable inferential structure that produces it. Within the Bulut Doctrine, narrative eect is theorized alon...

Assessment of Professional Medical Capabilities in Mainstream Chinese Large Language Models for Tremor-Related Diseases: A Comparative Study Based on Expert Scoring

APF Paper 10 — The Calculus of Finite Continuability

Paper 10 of the Admissibility Physics Framework (APF), The Calculus of Finite Continuability. Spencer-Brown's Laws of Form begins by drawing a distinction; APF begins one layer lower — a distinction is physical only through the admissible continuations it can sustain. Core results. The primitive judgment. Γ ⊢C D → E: i...

Silent Collapse of Ecosystems Revisited: Rate-Induced Tipping, Eco-Evolutionary Decoupling, and Multi-Layer Early Warning in the Anthropocene (LLM collaboration and Field Compute model August 2026 Update)

🌍 Silent Collapse of Ecosystems Revisited Rate-Induced Tipping · Eco-Evolutionary Decoupling · Multi-Layer Early Warning · Field Compute August–September 2026 Update This archive has evolved substantially since the first Silent Collapse formulation. What began as an attempt to describe the hidden erosion of ecological...

APF Paper 1 — The Enforceability of Distinction (Technical Supplement)

Technical Supplement to Paper 1 of the Admissibility Physics Framework (APF), The Enforceability of Distinction. The minimal defender quotient is a generated Boolean algebra modulo a kernel ideal. The process-theoretic defender is a partial Boolean-record functor. The supplement gives the LP-versus-NPA distinction in s...

Prompting Large Language Models with Partial Knowledge for Answering Questions with Unseen Entities

Retrieval-Augmented Generation (RAG) systems are often constrained by the “relevance myopia”, primarily focusing on fully relevant external information to supply missing or outdated facts while overlooking the potential of partially relevant knowledge to activate the internal knowledge of Large Language Models (LLMs)....

Guiding AI to Fix Its Own Flaws: An Empirical Study on LLM-Driven Secure Code Generation

Large Language Models have become powerful tools for programming. However, they often overlook essential security practices, producing code with vulnerabilities. Despite this risk, existing work offers limited guidance on steering LLMs toward secure code generation and lacks systematic analysis of how effectively LLMs...

FORK, RECOVERY, AND HISTORICALLY ORGANIZED CONTINUITY IN A LONGITUDINAL LLM–HUMAN INTERACTION: A Naturalistic Case Study

Long-running interactions with large language models can become historically organized, but apparent continuity across branches or reconstructed sessions may arise from several different carriers, including accessible history, memory systems, procedural state, lineage, or harness-level persistence. This naturalistic ca...

Toward Reproducible Local Evaluation of Prompt Injection in Open-Weight LLMs: Design and Threat Model

Open-weight language models are increasingly deployed in local and edge environments, yet their robustnessagainst prompt injection and system-prompt leakage remains poorly quantified under reproducible conditions.Existing large-scale benchmarks largely target hosted, API-served models and can be difficult to run end-to...

Toward Reproducible Local Evaluation of Prompt Injection in Open-Weight LLMs: Design and Threat Model

Open-weight language models are increasingly deployed in local and edge environments, yet their robustnessagainst prompt injection and system-prompt leakage remains poorly quantified under reproducible conditions.Existing large-scale benchmarks largely target hosted, API-served models and can be difficult to run end-to...

THE SANNER MODEL — APPLIED TO A FOUR-BODY SYSTEM

This paper applies The Sanner Model to a four-body system — the Galactic Center, the Sun, the Moon, and the Earth. Each mass acts as a sink for the aether, producing its own radial, inward aether acceleration field. The paper distinguishes sharply between two properties of the combined field: aether acceleration, which...