HFEPX Daily Archive: 2026-02-16

Research Utility Snapshot

Papers Published On This Date

• How to Train Your Long-Context Visual Document Model

  Austin Veselka · Feb 16, 2026

  Pairwise Preference

  We systematically study continued pretraining, supervised finetuning, and preference optimization for 24B and 32B parameter models, backed by extensive LC evaluations and ablations to bridge this gap, and achieve state-of-the-art performanc

• Colosseum: Auditing Collusion in Cooperative Multi-Agent Systems

  Mason Nakamura, Abhinav Kumar, Saswat Das, Sahar Abdelnabi, Saaduddin Mahmud · Feb 16, 2026

  Multi Agent

  Multi-agent systems, where LLM agents communicate through free-form language, enable sophisticated coordination for solving complex cooperative tasks.

• OpaqueToolsBench: Learning Nuances of Tool Behavior Through Interaction

  Skyler Hallinan, Thejas Venkatesh, Xiang Ren, Sai Praneeth Karimireddy, Ashwin Paranjape · Feb 16, 2026

  Tool Use

  Tool-calling is essential for Large Language Model (LLM) agents to complete real-world tasks.

• Weight space Detection of Backdoors in LoRA Adapters

  David Puertolas Merenciano, Ekaterina Vasyagina, Raghav Dixit, Kevin Zhu, Ruizhe Li · Feb 16, 2026

  We evaluate the method on 500 LoRA adapters -- 400 clean, and 100 poisoned for Llama-3.2-3B on instruction and reasoning datasets: Alpaca, Dolly, GSM8K, ARC-Challenge, SQuADv2, NaturalQuestions, HumanEval, and GLUE dataset.

• AIC CTU@AVerImaTeC: dual-retriever RAG for image-text fact checking

  Herbert Ullrich, Jan Drchal · Feb 16, 2026

  In this paper, we present our 3rd place system in the AVerImaTeC shared task, which combines our last year's retrieval-augmented generation (RAG) pipeline with a reverse image search (RIS) module.

• ScrapeGraphAI-100k: A Large-Scale Dataset for LLM-Based Web Information Extraction

  William Brach, Francesco Zuppichini, Marco Vinciguerra, Lorenzo Padoan · Feb 16, 2026

  ScrapeGraphAI-100k enables fine-tuning small models, benchmarking structured extraction, and studying schema induction for web IR indexing, and is publicly available on HuggingFace.

• Seeing to Generalize: How Visual Data Corrects Binding Shortcuts

  Nicolas Buzeta, Felipe del Rio, Cristian Hinostroza, Denis Parra, Hans Lobel · Feb 16, 2026

  Vision Language Models (VLMs) are designed to extend Large Language Models (LLMs) with visual capabilities, yet in this work we observe a surprising phenomenon: VLMs can outperform their underlying LLMs on purely text-only tasks, particular

• Protecting Language Models Against Unauthorized Distillation through Trace Rewriting

  Xinhang Ma, William Yeoh, Ning Zhang, Yevgeniy Vorobeychik · Feb 16, 2026

  Knowledge distillation is a widely adopted technique for transferring capabilities from LLMs to smaller, more efficient student models.

• CGRA-DeBERTa Concept Guided Residual Augmentation Transformer for Theologically Islamic Understanding

  Tahir Hussain, Saddam Hussain Khan · Feb 16, 2026

  The qualitative evaluation noted better extraction and discrimination and theological precision.

• Symmetry in language statistics shapes the geometry of model representations

  Dhruva Karkada, Daniel J. Korchinski, Andres Nava, Matthieu Wyart, Yasaman Bahri · Feb 16, 2026

  Although learned representations underlie neural networks' success, their fundamental properties remain poorly understood.

• Scaling Beyond Masked Diffusion Language Models

  Subham Sekhar Sahoo, Jean-Marie Lemercier, Zhihan Yang, Justin Deschenaux, Jingyu Liu · Feb 16, 2026

  Among discrete diffusion approaches, Masked diffusion currently dominates, largely driven by strong perplexity on language modeling benchmarks.

• Text Style Transfer with Parameter-efficient LLM Finetuning and Round-trip Translation

  Ruoxi Liu, Philipp Koehn · Feb 16, 2026

  This paper proposes a novel method for Text Style Transfer (TST) based on parameter-efficient fine-tuning of Large Language Models (LLMs).

• Cold-Start Personalization via Training-Free Priors from Structured World Models

  Avinandan Bose, Shuyue Stella Li, Faeze Brahman, Pang Wei Koh, Simon Shaolei Du · Feb 16, 2026

  Pairwise Preference

  Cold-start personalization requires inferring user preferences through interaction when no user-specific historical data is available.

• Learning User Interests via Reasoning and Distillation for Cross-Domain News Recommendation

  Mengdan Zhu, Yufan Zhao, Tao Di, Yulan Yan, Liang Zhao · Feb 16, 2026

  News recommendation plays a critical role in online news platforms by helping users discover relevant content.

• Counterfactual Fairness Evaluation of LLM-Based Contact Center Agent Quality Assurance System

  Kawin Mayilvaghanan, Siddhant Gupta, Ayush Kumar · Feb 16, 2026

  Large Language Models (LLMs) are increasingly deployed in contact-center Quality Assurance (QA) to automate agent performance evaluation and coaching feedback.

• Tool-Aware Planning in Contact Center AI: Evaluating LLMs through Lineage-Guided Query Decomposition

  Varun Nathan, Shreyas Guha, Ayush Kumar · Feb 16, 2026

  Critique Edit

  We present a domain-grounded framework and benchmark for tool-aware plan generation in contact centers, where answering a query for business insights, our target use case, requires decomposing it into executable steps over structured tools

• BFS-PO: Best-First Search for Large Reasoning Models

  Fiorenzo Parascandolo, Wenhui Tan, Enver Sangineto, Ruihua Song, Rita Cucchiara · Feb 16, 2026

  Using different benchmarks and base LRMs, we show that BFS-PO can simultaneously increase the LRM accuracy and shorten its answers.

• Testimole-Conversational: A 30-Billion-Word Italian Discussion Board Corpus (1996-2024) for Language Modeling and Sociolinguistic Research

  Matteo Rinaldi, Rossella Varvara, Viviana Patti · Feb 16, 2026

  We present "Testimole-conversational" a massive collection of discussion boards messages in the Italian language.

• Learning State-Tracking from Code Using Linear RNNs

  Julien Siems, Riccardo Grazzi, Kirill Kalinin, Hitesh Ballani, Babak Rahmani · Feb 16, 2026

  Over the last years, state-tracking tasks, particularly permutation composition, have become a testbed to understand the limits of sequence models architectures like Transformers and RNNs (linear and non-linear).

• Physical Commonsense Reasoning for Lower-Resourced Languages and Dialects: a Study on Basque

  Jaione Bengoetxea, Itziar Gonzalez-Dios, Rodrigo Agerri · Feb 16, 2026

  Physical commonsense reasoning represents a fundamental capability of human intelligence, enabling individuals to understand their environment, predict future events, and navigate physical spaces.

• Overthinking Loops in Agents: A Structural Risk via MCP Tools

  Yohan Lee, Jisoo Jang, Seoyeon Choi, Sangyeop Kim, Seungtaek Choi · Feb 16, 2026

  Tool-using LLM agents increasingly coordinate real workloads by selecting and chaining third-party tools based on text-visible metadata such as tool names, descriptions, and return messages.

• A Geometric Analysis of Small-sized Language Model Hallucinations

  Emanuele Ricco, Elia Onofri, Lorenzo Cima, Stefano Cresci, Roberto Di Pietro · Feb 16, 2026

  Long Horizon

  Hallucinations -- fluent but factually incorrect responses -- pose a major challenge to the reliability of language models, especially in multi-step or agentic settings.

• Emergently Misaligned Language Models Show Behavioral Self-Awareness That Shifts With Subsequent Realignment

  Laurène Vaugrante, Anietta Weckauff, Thilo Hagendorff · Feb 16, 2026

  Our findings show that behavioral self-awareness tracks actual alignment states of models, indicating that models can be queried for informative signals about their own safety.

• Multi-Agent Comedy Club: Investigating Community Discussion Effects on LLM Humor Generation

  Shiwei Hong, Lingyao Li, Ethan Z. Rong, Chenxinran Shen, Zhicong Lu · Feb 16, 2026

  Pairwise PreferenceRubric Rating Multi Agent

  Prior work has explored multi-turn interaction and feedback for LLM writing, but evaluations still largely center on prompts and localized feedback, leaving persistent public reception in online communities underexamined.

• Unlocking Reasoning Capability on Machine Translation in Large Language Models

  Sara Rajaee, Sebastian Vincent, Alexandre Berard, Marzieh Fadaee, Kelly Marchisio · Feb 16, 2026

  Critique Edit Long Horizon

  We systematically evaluate several open- and closed-weights RLMs on the WMT24++ benchmark and find that enabling explicit reasoning consistently degrades translation quality across languages and models.

• Residual Connections and the Causal Shift: Uncovering a Structural Misalignment in Transformers

  Jonathan Lys, Vincent Gripon, Bastien Pasdeloup, Lukas Mauch, Fabien Cardinaux · Feb 16, 2026

  Experiments on multiple benchmarks show that these strategies alleviate the representation misalignment and yield improvements, providing an efficient and general architectural enhancement for autoregressive Transformers.

• Cognitive networks reconstruct mindsets about STEM subjects and educational contexts in almost 1000 high-schoolers, University students and LLM-based digital twins

  Francesco Gariboldi, Emma Franchino, Edith Haim, Gianluca Lattanzi, Alessandro Grecucci · Feb 16, 2026

  Human networks show greater overlapping between mathematics and anxiety than GPT-oss.

• Rethinking the Role of LLMs in Time Series Forecasting

  Xin Qiu, Junlong Tong, Yirong Sun, Yunpu Ma, Wei Zhang · Feb 16, 2026

  We show that such conclusions stem from limited evaluation settings and do not hold at scale.

• LLMStructBench: Benchmarking Large Language Model Structured Data Extraction

  Sönke Tenckhoff, Mario Koddenbrock, Erik Rodner · Feb 16, 2026

  We present LLMStructBench, a novel benchmark for evaluating Large Language Models (LLMs) on extracting structured data and generating valid JavaScript Object Notation (JSON) outputs from natural-language text.

• Evolutionary System Prompt Learning for Reinforcement Learning in LLMs

  Lunjun Zhang, Ryan Chen, Bradly C. Stadie · Feb 16, 2026

  Building agentic systems that can autonomously self-improve from experience is a longstanding goal of AI.

• Exposing the Systematic Vulnerability of Open-Weight Models to Prefill Attacks

  Lukas Struppek, Adam Gleave, Kellin Pelrine · Feb 16, 2026

  Red Team

  As the capabilities of large language models continue to advance, so does their potential for misuse.

• Crowdsourcing Piedmontese to Test LLMs on Non-Standard Orthography

  Gianluca Vico, Jindřich Libovický · Feb 16, 2026

  We use this resource to benchmark several large language models on tokenization parity, topic classification, and machine translation.

• Breaking Data Efficiency Dilemma: A Federated and Augmented Learning Framework For Alzheimer's Disease Detection via Speech

  Xiao Wei, Bin Wen, Yuqin Lin, Kai Li, Mingyang gu · Feb 16, 2026

  Early diagnosis of Alzheimer's Disease (AD) is crucial for delaying its progression.

• Is Information Density Uniform when Utterances are Grounded on Perception and Discourse?

  Matteo Gay, Coleman Haley, Mario Giulianelli, Edoardo Ponti · Feb 16, 2026

  The Uniform Information Density (UID) hypothesis posits that speakers are subject to a communicative pressure to distribute information evenly within utterances, minimising surprisal variance.

• Query as Anchor: Scenario-Adaptive User Representation via Large Language Model

  Jiahao Yuan, Yike Xu, Jinyong Wen, Baokun Wang, Ziyi Gao · Feb 16, 2026

  Evaluations on 10 Alipay industrial benchmarks show consistent SOTA performance, strong scalability, and efficient deployment.

• BETA-Labeling for Multilingual Dataset Construction in Low-Resource IR

  Md. Najib Hasan, Mst. Jannatun Ferdous Rain, Fyad Mohammed, Nazmul Siddique · Feb 16, 2026

  Manual annotation is expensive and difficult to scale, while using large language models (LLMs) as automated annotators introduces concerns about label reliability, bias, and evaluation validity.

Recent Daily Archives