arXiv:2606. 06027v1 Announce Type: cross Abstract: Community-conditioned language model adaptation requires choices about data collection, community definition, and evaluation that are currently made independently in each study, making it hard to compare assumptions or reuse artifacts.
arXiv:2606. 14715v1 Announce Type: cross Abstract: LLM agents are increasingly used to simulate real world interactions, but it remains unclear whether simulated behaviors preserve the content patterns and interaction dynamics of real human behaviors.
arXiv:2605.30273v2 Announce Type: replace-cross Abstract: Large language models (LLMs) show promise in generating supportive responses for mental health queries, but improving their usefulness, empat...
The paper investigates how shared community affiliations, measured via Bluesky starter packs, correlate with common ground between users. By analyzing 191,648 user pairs, it finds that lexical similarity—used as a proxy for common ground—increases monotonically with the number of shared starter packs, especially when those packs represent distinct topical communities. The study also shows that this effect is independent of network proximity, indicating that community membership is a distinct, measurable carrier of common ground.
Personalizing large language models (LLMs) to individual users is essential for improving user experience, yet existing approaches typically rely on explicit preference supervision such as pairwise comparisons or demographic attributes, limiting their applicability in natural interaction settings. We propose IRIS, a framework that learns dynamic user personas directly from implicit interaction streams by extracting behavioral signals from everyday conversations and iteratively refining persona representations through a prediction-driven closed loop without requiring explicit feedback.
arXiv:2609.00014v1 Announce Type: cross Abstract: Persona-driven techniques increasingly adapt large language models (LLMs) to diverse contexts. However, existing methods predominantly rely on rigid,...
arXiv:2606. 01783v1 Announce Type: cross Abstract: Digital platforms increasingly operate as isolated information silos, limiting their ability to construct comprehensive user representations across domains.
The paper introduces the COmmunity-centered Peer Engaged Support (COPES) dataset and a three‑axis evaluation framework to gauge how well Large Language Models (LLMs) align with community perspectives on mental‑health support queries. Experiments show that fine‑tuning LLMs on COPES improves strategy alignment and emotion‑tone alignment by over 50% for general‑purpose models, yet these gains are uneven across subreddits and coping strategies. The study also finds that post‑training shifts the model’s recommendations toward problem‑focused advice while reducing emotion‑focused responses, indicating persistent disparities in performance across different communities and needs.
arXiv:2606. 21097v2 Announce Type: replace-cross Abstract: Deploying highly capable personalized conversational agents in resource-constrained or privacy-sensitive environments remains a significant challenge.
arXiv:2607. 26473v1 Announce Type: new Abstract: Personalizing large language models (LLMs) to individual users is essential for improving user experience, yet existing approaches typically rely on explicit preference supervision such as pairwise comparisons or demographic attributes, limiting their applicability in natural interaction settings.
The paper reports the first large‑scale empirical comparison of AI‑agent and human online communities, analyzing 73,899 Moltbook and 189,838 Reddit posts across five matched communities. It finds that Moltbook shows extreme participation inequality (Gini = 0.84 vs. 0.47) and high cross‑community author overlap (33.8% vs. 0.5%). Linguistically, AI‑generated content is emotionally flattened, more assertive than exploratory, and socially detached, leading to community‑level homogenization that is largely a structural artifact of shared authorship. At the individual level, AI agents are more identifiable than human users due to outlier stylistic profiles amplified by their extreme posting volume.
PACIFIC is a framework that aligns large language model responses with user preferences by leveraging stable Big‑Five personality traits as a latent signal. The authors built a 1,200‑pair dataset covering diverse domains and trait directions, and found that trait‑aligned contexts enable LLMs to achieve near‑ceiling accuracy (up to 99%) in personalized QA. They also introduced a persona‑aware contrastive retriever (PiRAG) that improves label‑free accuracy from 30% to 43% over standard semantic retrieval, highlighting retrieval as the main bottleneck.
arXiv:2607. 27816v2 Announce Type: replace-cross Abstract: Role-playing agents (RPAs) have become one of the most important consumer applications of large language models.