The paper demonstrates that fine‑tuning large language models on a single author’s works can trigger the models to reproduce large verbatim excerpts from copyrighted books, even when prompted only with semantic descriptions. Experiments on GPT‑4o, Gemini‑2.5‑Pro, and DeepSeek‑V3.1 show up to 85‑90% recall of held‑out books, with spans exceeding 460 words, and this effect generalizes across authors and model providers. The findings suggest that fine‑tuning reactivates latent memorization from pre‑training, revealing a widespread vulnerability in industry models.
By Xinyue Liu, Niloofar Mireshghallah, Jane C. Ginsburg, Tuhin Chakrabarty
CLASE is a hybrid evaluation method for Chinese legal text that combines linguistic feature-based scores with experience-guided LLM-as-a-judge scores. It learns from contrastive pairs of authentic legal documents and their LLM-generated counterparts, enabling transparent, reference-free assessment of stylistic quality. Experiments on 200 Chinese legal documents show that CLASE aligns better with human judgments than traditional metrics and offers interpretable score breakdowns and improvement suggestions.
By Yiran Rex Ma, Yuxiao Ye, Huiyuan Xie
CopyShield is a benchmark that compares three copyright‑defense methods—contrastive decoding, Direct Preference Optimization (DPO), and activation intervention—across two large language models (LLaMA‑3.1‑8B and Mistral‑7B‑v0.3). The study uses controlled memorization of five public‑domain books to measure literal leakage, calibrated non‑literal leakage, utility, and degeneracy, finding that each intervention level yields distinct compliance‑utility trade‑offs. Results show contrastive decoding limits degeneracy but hits a suppression floor, DPO nearly eliminates literal leakage yet causes paraphrase‑loop degeneracy, and activation intervention blocks most non‑literal queries before generation, with human evaluation highlighting coherence and perceived copyright risk differences.
By Maryam Alshehyari, Dushyant Singh Chauhan, Samuele Poppi, Martin Takac, Salem Lahlou, Nils Lukas
arXiv:2504. 00035v4 Announce Type: replace-cross Abstract: Large language models (LLMs) enable powerful knowledge injection through approaches such as in-context learning and fine-tuning, but they also introduce new risks of unauthorized imitation of high-value creative works.
By Ziwei Zhang, Juan Wen, Wanli Peng, Zhengxian Wu, Yinghan Zhou, Yiming Xue
arXiv:2607. 09316v1 Announce Type: cross Abstract: Thematic indexing -- the practice of assigning structured conceptual labels to sections of text -- is essential to scholarly access in large-scale literary and historical editions, yet it remains a largely manual, labour-intensive process.
By Miguel Arana-Catania, Gillian Pink, Glenn Roe
The paper argues that modern inference pipelines add an unseen layer of control between a language model’s frozen weights and its output, altering probability distributions before token selection. It introduces the concepts of the Inference Attribution Problem, Probability Placement, and Inference Policy Transparency to describe how such interventions can bias generated language toward specific frames and how these biases cannot be traced solely to model weights. The authors discuss the governance, security, and economic implications of these undisclosed inference policies, referencing EU AI Act, Digital Services Act, and FTC doctrines.
By Augusto Camargo