LiSeCo is a lightweight, gradient‑free method that controls language generation by directly intervening on the hidden activations of a token in embedding space. It uses control‑theoretic techniques to steer the generation trajectory away from undesired semantic regions and into a predefined allowed region, ensuring fine‑grained attribute control. The approach is computationally efficient, minimally impacts generation time, and is shown to be effective on tasks such as toxicity, sentiment, and bilingual language steering while preserving text quality.
By Emily Cheng, Carmen Amo Alonso
The paper introduces KESA, a knowledge‑enhanced approach for sentence‑level sentiment analysis that incorporates sentiment knowledge through two auxiliary tasks: sentiment word cloze and conditional sentiment prediction. These tasks use prior sentiment polarity to guide the selection of sentiment words and the prediction of overall sentiment, respectively, and explore label combination methods to unify multiple label types. Experiments show that KESA consistently outperforms pre‑trained models and complements existing knowledge‑enhanced post‑training methods.
By Qinghua Zhao, Shuai Ma, Shuo Ren
The paper introduces Retrieval-Augmented Decoding (RAD), a decoding-time method that improves the truthfulness of large language models without retraining. RAD uses a small reference set of up to ten annotated examples to build a grounding space of context embeddings and next-token logits, which it retrieves and aggregates during inference to shape the model’s output. Experiments on four open-ended generation benchmarks and four different LLMs show that RAD consistently outperforms strong baselines and generalizes well across tasks.
By Manh Nguyen, Sunil Gupta, Hung Le
arXiv:2606.18389v2 Announce Type: replace
Abstract: Large language models (LLMs) have become an effective tool for synthetic data generation, including for low-resource languages, where generated dat...
By Jan Cegin, Daniil Gurgurov, Yusser Al Ghussin, Simon Ostermann
arXiv:2606. 08408v1 Announce Type: cross Abstract: We extend activation steering to diffusion language models (DLMs) and study a novel problem that arose due to the inference mechanism of DLMs: Modifying a text in-place to manifest a different concept.
By Ryandito Diandaru, Ikhlasul Akmal Hanif, Fadli Aulawi Al Ghiffari, Ahmed Elshabrawy, Alham Fikri Aji
arXiv:2607. 17524v1 Announce Type: cross Abstract: We propose Token-Level Off-Policy Labeling (TOPL), an off-policy training paradigm that reframes post-training as a token-level correctness prediction task.
By Zitong Huang, Gustavo Lucas Carvalho, Deqing Fu, Robin Jia
The paper introduces a syntax-guided diffusion language model that incorporates structural supervision and personalized conditioning to improve text quality, diversity, and controllability. It presents a cascaded framework generating syntactic guidance before text generation, and a novel noncascaded architecture for better structure-content alignment. A shared representation mechanism enables fine‑grained personalization across users, achieving faithful stylistic generation and zero‑shot inference, with experiments showing superior fluency, diversity, and stylistic fidelity.
By Ruqian Zhang, Yijiao Zhang, Juan Shen, Zhongyi Zhu, Annie Qu
arXiv:2601.21647v2 Announce Type: replace-cross
Abstract: Discrete Diffusion Language Models (DLMs) offer a promising non-autoregressive alternative for text generation, yet effective mechanisms for...
By Eden Avrahami, Eliya Nachmani
The paper introduces a dataset watermarking technique that embeds a watermark by increasing the co‑occurrence of randomly selected word pairs through meaning‑preserving local edits. The watermark can be detected solely from generated text with provable false‑positive control, and experiments on four base models and three datasets show reliable detection (p < 0.01) even when the watermarked data constitutes less than 5% of fine‑tuning tokens. Compared to existing methods, the approach better preserves benchmark utility and semantic integrity.
By Pengrun Huang, Kamalika Chaudhuri, Yu-Xiang Wang
arXiv:2608.30425v1 Announce Type: new
Abstract: Cross-lingual aspect-based sentiment analysis (ABSA) transfers knowledge from a source language with annotated data to a target language, enabling fine...
By Jakub \v{S}m\'{i}d, Pavel P\v{r}ib\'{a}\v{n}, Pavel Kr\'{a}l
The paper introduces a steering‑vector‑based causal attribution framework to study how large vision‑language models (LVLMs) translate visual input into emotional narratives. By creating a specialized dataset, the authors uncover a functional decoupling in the LVLM’s three‑stage Adapt‑Aggregate‑Execute mechanism: visual emotional cues are first aggregated in middle layers via sentiment‑specific attention heads, then translated into narrative generation in deeper layers through emotion‑general pathways. Using these insights, they regulate emotional information routing to strengthen attention flow and amplify semantic activation, achieving significant performance gains on the MER‑UniBench and reducing emotional hallucinations through inference‑time intervention.
By Chengsheng Zhang, Chenghao Sun, Zhining Xie, Xinmei Tian
arXiv:2606.28249v2 Announce Type: replace-cross
Abstract: Recently, Large Language Model (LLM)-based Text-to-Speech (TTS) models have achieved remarkable naturalness. However, the standard Supervised...
By Sihang Nie, Xiaofen Xing, Rui Xing, Haoming Li, Ruitong Xiao, Jingyuan Xing, Baiji Liu, Xiangmin Xu