arXiv subjects

Joshua Freeman

Publications and source records attributed to Joshua Freeman.

3 recordsLinked to original sources

arXiv2026

Second harmonic generation and third harmonic generation in topological insulator-based van der Waals metamaterials

High-order harmonic generation (HHG) in solids - the frequency up-conversion of an optical signal - is governed by symmetries. At terahertz (THz) frequencies, HHG is a key technology to access high frequency spectral windows that are usually difficult to cover using conventional solid state laser technologies. This effect has been recently exploited in graphene where HHG has been demonstrated, albeit only at odd multiples of the driving frequency owing to its inherent centro-symmetry. In topological insulators (TIs), the combination of spin-orbit interaction and time-reversal symmetry create an insulating bulk state with an inverted band order, inseparably connected with conducting surface states. TIs have been predicted to support unconventional high harmonic generation from the bulk and topological surface, which are usually difficult to be distinguished. However, no experimental results have been provided, so far. Here, we exploit the strong optical field amplification provided by an array of single or double split ring resonators, with embedded Bi2Se3 or (InxBi(1-x))2Se3/ Bi2Se3 van der Waals heterostructures, to achieve up-conversion in the 6.4 (even) - 9.7 (odd) THz frequency range. This results from bulk centro-symmetry (odd states) and symmetry breaking in the topological surface states (odd and even).

physics.optics↗

arXiv2025

Real-Time Detection of Hallucinated Entities in Long-Form Generation

Large language models are now routinely used in high-stakes applications where hallucinations can cause serious harm, such as medical consultations or legal advice. Existing hallucination detection methods, however, are impractical for real-world use, as they are either limited to short factual queries or require costly external verification. We present a cheap, scalable method for real-time identification of hallucinated tokens in long-form generations, and scale it effectively to 70B parameter models. Our approach targets entity-level hallucinations-e.g., fabricated names, dates, citations-rather than claim-level, thereby naturally mapping to token-level labels and enabling streaming detection. We develop an annotation methodology that leverages web search to annotate model responses with grounded labels indicating which tokens correspond to fabricated entities. This dataset enables us to train effective hallucination classifiers with simple and efficient methods such as linear probes. Evaluating across four model families, our classifiers consistently outperform baselines on long-form responses, including more expensive methods such as semantic entropy (e.g., AUC 0.90 vs 0.71 for Llama-3.3-70B), and are also an improvement in short-form question-answering settings. Despite being trained only to detect hallucinated entities, our probes effectively detect incorrect answers in mathematical reasoning tasks, indicating generalization beyond entities. While our annotation methodology is expensive, we find that annotated responses from one model can be used to train effective classifiers on other models; accordingly, we publicly release our datasets to facilitate reuse. Overall, our work suggests a promising new approach for scalable, real-world hallucination detection.

cs.CL↗

arXiv2024

Exploring Memorization and Copyright Violation in Frontier LLMs: A Study of the New York Times v. OpenAI 2023 Lawsuit

Copyright infringement in frontier LLMs has received much attention recently due to the New York Times v. OpenAI lawsuit, filed in December 2023. The New York Times claims that GPT-4 has infringed its copyrights by reproducing articles for use in LLM training and by memorizing the inputs, thereby publicly displaying them in LLM outputs. Our work aims to measure the propensity of OpenAI's LLMs to exhibit verbatim memorization in its outputs relative to other LLMs, specifically focusing on news articles. We discover that both GPT and Claude models use refusal training and output filters to prevent verbatim output of the memorized articles. We apply a basic prompt template to bypass the refusal training and show that OpenAI models are currently less prone to memorization elicitation than models from Meta, Mistral, and Anthropic. We find that as models increase in size, especially beyond 100 billion parameters, they demonstrate significantly greater capacity for memorization. Our findings have practical implications for training: more attention must be placed on preventing verbatim memorization in very large models. Our findings also have legal significance: in assessing the relative memorization capacity of OpenAI's LLMs, we probe the strength of The New York Times's copyright infringement claims and OpenAI's legal defenses, while underscoring issues at the intersection of generative AI, law, and policy.

cs.LG↗