This paper investigates whether a group of people highlighting the same document forms a single consensus or is internally structured into reader sub-groups.

The paper introduces a novel, per-token feature derived from how sampling temperature reshapes the token distribution, demonstrating it is a significantly stronger predictor of LLM creativity than sta…

The paper proposes a comprehensive benchmark to systematically audit how varying persona prompts and model choices affect the technical quality and social representativeness of scholar recommendations…

The paper demonstrates that the AI-like style introduced by post-training alignment can be measured, localized, and causally removed using a novel ablation technique called PASTA.

The paper introduces a new benchmark for multi-target cross-lingual summarization (MTXLS) and proposes an activation steering method that significantly improves LLM performance by guiding the generati…

The paper introduces a Deep Research pipeline that significantly improves literature search recall and demonstrates that human-curated citation lists are often unreliable and do not serve as a true gr…

This paper conducts a large-scale audit of human annotation reporting in NLP, finding that while reporting has improved, critical details needed to assess annotation validity, such as training and agr…

The study demonstrates that conditioning AI brand recommendations on a user's persona significantly alters the recommended product set, particularly for mid-market brands, and this effect is largest o…

The paper demonstrates that supervised fine-tuning significantly outperforms frontier zero-shot large language models for screen-conditioned action prediction on the PiSAR benchmark, highlighting the…

The paper introduces TELL, a novel explainable AI-generated text detection architecture that provides detailed, human-understandable explanations for its scores, achieving competitive performance whil…

The paper demonstrates that LLM performance in zero-shot annotation is significantly limited by the alignment between the model's internal understanding and the task definition, showing that prompt-ba…

The paper introduces BiAxisAudit, a novel framework that evaluates LLM bias by analyzing bias scores across multiple prompt formats and within the internal inconsistency of model responses, revealing…

SkillPager is a novel two-stage framework that efficiently selects minimal, execution-sufficient context from large procedural skill documents by leveraging typed semantic nodes, significantly reducin…

The paper proposes 'Uncertainty,' a multiscale uncertainty estimator that focuses on low-probability tokens to improve the detection of AI-generated text by addressing boilerplate dominance and score…

The paper introduces OpAI-Bench, a novel benchmark designed to study how AI authorship signals evolve and accumulate during the progressive co-editing process between humans and AI.

The paper introduces SmartIterator (SI), a visual analytics framework that systematically guides analysts through the complex process of evaluating and understanding how data groupings change across p…

Xetrieval introduces an embedding-level framework to mechanistically explain dense retrieval decisions by decomposing high-dimensional embeddings into sparse, human-interpretable features.

The paper introduces Dr. DocBench, a difficulty-aware, comprehensive benchmark designed to rigorously test expert-level and challenging document parsing capabilities for VLMs, demonstrating that curre…

The paper introduces the Decan metric, a novel, information-theoretic approach for measuring creative diversity in AI outputs, which successfully detects diversity loss across different model fine-tun…

The paper introduces TorchSight, an open-source local system using a fine-tuned Qwen 3.5 27B model that achieves high accuracy (95.0%) in classifying sensitive security documents without relying on ex…