In many human-annotated NLP tasks involving ambiguity or subjective judgment, annotator disagreement reflects epistemic uncertainty rather than noise. Soft labeling (SL), which represents annotations as probability distributions rather than majority-vote (MV) labels, preserves this uncertainty and can improve downstream performance. We extend this perspective to LLM-based annotation by formalizing LLM soft labeling as introducing controlled variation in model-generated annotations to approximate the latent variability underlying human annotations. We distinguish two sources of variation: model-induced (e.g., stochastic decoding and model ensembles) and human-approximated (e.g., persona prompting and human-calibrated in-context annotation). Using the Gab Hate and GoEmotions datasets, we show that SL training consistently outperforms MV training under stronger LLM-based annotation strategies. Model ensembles produce the most informative soft-label distributions, achieving the best human–LLM agreement and downstream classification performance. These findings suggest that scalable LLM-based annotation pipelines can model epistemic uncertainty through diverse model-level variation without explicitly simulating human attributes.

Humans develop cooperation heuristics in social decision-making, either intuitively or deliberatively. Large language models (LLMs), which exhibit human-like heuristics across cognitive domains, may acquire prosocial tendencies through instruction tuning, latently encoded in their representations to foster cooperative behavior in social reasoning games. However, most studies of this kind either focus on cooperative language generation or explicitly instruct LLMs to cooperate, deviating from the inherent cooperation heuristics of humans. Our negotiation role-play simulations using BATNA (Best Alternative to a Negotiated Agreement) with a GPT-based LLM reveal that LLMs may struggle with cooperation in the absence of explicit instructions, showing a 50–90% lower success rate than in instructed scenarios and a 40–80% lower success rate than human performance reported in past studies. Implicitly inducing cooperation through personality traits had inconsistent effects, with agreeableness showing only a marginal influence and other traits exhibiting no systematic impact. These findings suggest that personality-based cooperation cues are subtle and that explicit instructions may still be essential for multi-agent LLMs to approximate human-like negotiation.

Media framing refers to highlighting certain aspect of an issue in the news to promote a particular interpretation to the audience. Supervised learning has often been used to recognize frames in news articles, requiring a known pool of frames for a particular issue, which must be identified by communication researchers through thorough manual content analysis. In this work, we devise an unsupervised learning approach to discover the frames in news articles automatically. Given a set of news articles for a given issue, e.g., gun violence, our method first extracts frame elements from these articles using related Wikipedia articles and the Wikipedia category system. It then uses a community detection approach to identify frames from these frame elements. We discuss the effectiveness of our approach by comparing the frames it generates in an unsupervised manner to the domain-expert-derived frames for the issue of gun violence, for which a supervised learning model for frame recognition exists.