First stable release of the toolkit that tests the structural hallucination thesis directly: that a language model's confidence tracks consistency with its own training corpus, not correspondence with the world. A small causal LM (Qwen2.5-0.5B by default) is fine-tuned on a corpus that consistently asserts a fabricated...
José Luciano Verçosa Marques, Frederico Jorge Heitmann, Daniel Omar Pérez et al.· Zenodo (CERN European Organi...· 0 citations
This paper reports the first executed computational proof-of-concept and cross-corpus replication of Bacterialized Gematria, a population-based symbol–number system in which an initially fixed A=1 through Z=26 correspondence table is transformed into a historically contingent numerical ecology capable of mutation, hori...
Kenji Siratori· Zenodo (CERN European Organi...· 0 citations
First stable release of the toolkit that tests the structural hallucination thesis directly: that a language model's confidence tracks consistency with its own training corpus, not correspondence with the world. A small causal LM (Qwen2.5-0.5B by default) is fine-tuned on a corpus that consistently asserts a fabricated...
José Luciano Verçosa Marques, Frederico Jorge Heitmann, Daniel Omar Pérez et al.· Zenodo (CERN European Organi...· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
Preprint, version 1.0 (September 2026) The Transformer of 2017 still sits at the core of every current large language model, yet almost everything around it has been rebuilt, some parts repeatedly. This survey traces the architectural development of text-only LLMs from 2017 to early 2026 as four epochs, each driven by...
Finn Reimann· Zenodo (CERN European Organi...· 0 citations
Yin‑Yang Fractal Unified Field Theory: Original 30‑volume monograph archive (Version 1) This submission serves as a permanent archival deposit for the original complete manuscripts of the monograph Yin‑Yang Fractal Unified Field Theory.The former DOI assigned to earlier uploads is no longer valid; this record provides...
Shuping Li· Zenodo (CERN European Organi...· 0 citations
Large language models can readily produce diverse and unusual outputs, yet increased variation does not by itself explain when a system becomes more capable of creative thought. This paper develops a preliminary hypothesis that prompt design may alter the effective dynamical regime of inference and, under some conditio...
Abstract In this feasibility study, we sought to explore how the Sketch and Speak intervention could be provided in a small group setting with a focus on main ideas and details to better mimic “real world” services provided in schools. Two fifth-grade participants completed eight sessions of intervention and pre-/postt...
Amy K. Peterson, Shiloh R. Gregory· Seminars in Speech and Langu...· 0 citations
OCR for low-resource languages is still held back by the same small number of issues: too little labeled image-text data, too few benchmarks, and thin language-specific tooling. Kurdish Sorani is a particularly awkward case. It is written in a modified Arabic script, runs right to left, and has orthographic habits that...
StudentSim is a research system for evaluating and improving AI tutoring tools before they are tested with learners. It models how individuals may engage with tutors, giving feedback to improve tutor performance across subjects like chess, writing, and mathematics. The post Simulating student learning for better AI tutoring appeared first on Microsoft Research.
