搜索结果：responses

Many in-silico simulations of human survey responses with large language models (LLMs) focus on generating closed-ended survey responses, whereas LLMs are typically trained to generate open-ended text instead. Previous research has used a diverse range of methods for generating closed-ended survey responses with LLMs, and a standard practice remains to be identified. In this paper, we systematically investigate the impact that various Survey Response Generation Methods have on predicted survey responses. We present the results of 32 mio. simulated survey responses across 8 Survey Response Generation Methods, 4 political attitude surveys, and 10 open-weight language models. We find significant differences between the Survey Response Generation Methods in both individual-level and subpopulation-level alignment. Our results show that Restricted Generation Methods perform best overall, and that reasoning output does not consistently improve alignment. Our work underlines the significant impact that Survey Response Generation Methods have on simulated survey responses, and we develop practical recommendations on the application of Survey Response Generation Methods.

COVID 19 pandemic has disrupted the global market and workplace landscape. As a response, hybrid work situations have become popular in the software business sector. This way of working has an impact on software companies. This study investigates software companies responses to hybrid working. We conducted a large scale survey to achieve our objective. Our results are based on a qualitative analysis of 124 valid responses. The main result of our study is a taxonomy of software companies impacts on hybrid working at individual, team and organisation levels. We found higher positive responses at individual and organisational levels than negative responses. At the team level, both positive and negative impacts obtained a uniform number of responses. The results indicate that hybrid working became credible with the wave of COVID 19, with 83 positive responses outweighing the 41 negative responses. Software company respondents witnessed better work-life balance, productivity, and efficiency in hybrid working.

Large Language Models are increasingly deployed in emotional-support contexts and crisis-related situations. Nevertheless, their cross-lingual abilities in these circumstances remain underexplored. Existing benchmarks emphasize multilingual performance but rarely examine crisis-related empathy and cultural grounding in low-to-mid-resource languages. We introduce SPLIT, a 500-prompt benchmark designed to evaluate LLM consistency in generating emotionally grounded responses across five categories: Stress, Panic, Loneliness, Internal Displacement, and Tension. We evaluate three technically diverse LLMs across three dimensions: Empathetic Accuracy, Linguistic Naturalness, and Contextual & Cultural Grounding. The framework aims to assess and compare the quality of LLM responses in both English and Ukrainian languages, as well as to explore the reliability of the LLM-as-a-jury paradigm. Our findings reveal that Gemini-2.5-Flash and LLaMA-3.3-70B-Instruct degrade when transitioning to Ukrainian, while DeepSeek-V3 remains comparatively stable within our benchmark. We additionally find that human and AI evaluators agree weakly on empathy and naturalness but diverge on cultural grounding

Large Language Model (LLM)-based conversational agents offer promising solutions for mental health support, but lack cultural responsiveness for diverse populations. This study evaluated the effectiveness of cultural prompting in improving cultural responsiveness and perceived empathy of LLM-generated therapeutic responses for Chinese American family caregivers. Using a randomized controlled experiment, we compared GPT-4o and Deepseek-V3 responses with and without cultural prompting. Thirty-six participants evaluated input-response pairs on cultural responsiveness (competence and relevance) and perceived empathy. Results showed that cultural prompting significantly enhanced GPT-4o's performance across all dimensions, with GPT-4o with cultural prompting being the most preferred, while improvements in DeepSeek-V3 responses were not significant. Mediation analysis revealed that cultural prompting improved empathy through improving cultural responsiveness. This study demonstrated that prompt-based techniques can effectively enhance the cultural responsiveness of LLM-generated therapeutic responses, highlighting the importance of cultural responsiveness in delivering empathetic AI-based

In the absence of an authoritative statement about a rumor, people may expose the truth behind such rumor through their responses on social media. Most rumor detection methods aggregate the information of all the responses and have made great progress. However, due to the different backgrounds of users, the responses have different relevance for discovering th suspicious points hidden in a rumor claim. The methods that focus on all the responding tweets would dilute the effect of the critical ones. Moreover, for a multi-modal rumor claim, the focus of a user may be on several words in the text or an object in the image, so the different modalities should be considered to select the relevant responses and verify the claim. In this paper, we propose a novel multi-modal rumor detection model, termed Focal Reasoning Model (FoRM), to filter out the irrelevant responses and further conduct fine-grained reasoning with the multi-modal claim and corresponding responses. Concretely, there are two main components in our FoRM: the coarse-grained selection and the fine-grained reasoning. The coarse-grained selection component leverages the post-level features of the responses to verify the clai

End-to-end (E2E) task-oriented dialogue (ToD) systems are prone to fall into the so-called "likelihood trap", resulting in generated responses which are dull, repetitive, and often inconsistent with dialogue history. Comparing ranked lists of multiple generated responses against the "gold response" (from evaluation data) reveals a wide diversity in response quality, with many good responses placed lower in the ranked list. The main challenge, addressed in this work, is then how to reach beyond greedily generated system responses, that is, how to obtain and select such high-quality responses from the list of overgenerated responses at inference without availability of the gold response. To this end, we propose a simple yet effective reranking method which aims to select high-quality items from the lists of responses initially overgenerated by the system. The idea is to use any sequence-level (similarity) scoring function to divide the semantic space of responses into high-scoring versus low-scoring partitions. At training, the high-scoring partition comprises all generated responses whose similarity to the gold response is higher than the similarity of the greedy response to the gol

Recent developments in natural language processing have demonstrated the potential of large language models (LLMs) to improve a range of educational and learning outcomes. Of recent chatbots based on LLMs, ChatGPT and Bard have made it clear that artificial intelligence (AI) technology will have significant implications on the way we obtain and search for information. However, these tools sometimes produce text that is convincing, but often incorrect, known as hallucinations. As such, their use can distort scientific facts and spread misinformation. To counter polarizing responses on these tools, it is critical to provide an overview of such responses so stakeholders can determine which topics tend to produce more contentious responses -- key to developing targeted regulatory policy and interventions. In addition, there currently exists no annotated dataset of ChatGPT and Bard responses around possibly polarizing topics, central to the above aims. We address the indicated issues through the following contribution: Focusing on highly polarizing topics in the US, we created and described a dataset of ChatGPT and Bard responses. Broadly, our results indicated a left-leaning bias for b

Colour is a fundamental determinant of affective experience in immersive virtual reality (VR), yet the emotional and physiological impact of individual hues remains poorly characterised. This study investigated how fifteen calibrated Munsell hues influence subjective and autonomic responses when presented in immersive VR. Thirty-six adults (18-45 years) viewed each hue in a within-subject design while pupil diameter and skin conductance were recorded continuously, and self-reported emotions were assessed using the Self-Assessment Manikin across pleasure, arousal, and dominance. Repeated-measures ANOVAs revealed robust hue effects on all three self-report dimensions and on pupil dilation, with medium to large effect sizes. Reds and red-purple hues elicited the highest arousal and dominance, whereas blue-green hues were rated most pleasurable. Pupil dilation closely tracked arousal ratings, while skin conductance showed no reliable hue differentiation, likely due to the brief (30 s) exposures. Individual differences in cognitive style and personality modulated overall reactivity but did not alter the relative ranking of hues. Taken together, these findings provide the first systemati

Social interactions promote well-being, yet barriers like geographic distance, time limitations, and mental health conditions can limit face-to-face interactions. Emotionally responsive AI systems, such as chatbots, offer new opportunities for social and emotional support, but raise critical questions about how empathy is perceived and experienced in human-AI interactions. This study examines how empathy is evaluated in AI-generated versus human responses. Using personal narratives, we explored how persona attributes (e.g., gender, empathic traits, shared experiences) and story qualities affect empathy ratings. We compared responses from standard and fine-tuned AI models with human judgments. Results show that while humans are highly sensitive to emotional vividness and shared experience, AI-responses are less influenced by these cues, often lack nuance in empathic expression. These findings highlight challenges in designing emotionally intelligent systems that respond meaningfully across diverse users and contexts, and informs the design of ethically aware tools to support social connection and well-being.

Reinforcement Learning from Human Feedback (RLHF) facilitates the alignment of large language models with human preferences, significantly enhancing the quality of interactions between humans and models. InstructGPT implements RLHF through several stages, including Supervised Fine-Tuning (SFT), reward model training, and Proximal Policy Optimization (PPO). However, PPO is sensitive to hyperparameters and requires multiple models in its standard implementation, making it hard to train and scale up to larger parameter counts. In contrast, we propose a novel learning paradigm called RRHF, which scores sampled responses from different sources via a logarithm of conditional probabilities and learns to align these probabilities with human preferences through ranking loss. RRHF can leverage sampled responses from various sources including the model responses from itself, other large language model responses, and human expert responses to learn to rank them. RRHF only needs 1 to 2 models during tuning and can efficiently align language models with human preferences robustly without complex hyperparameter tuning. Additionally, RRHF can be considered an extension of SFT and reward model trai

Low frequency seismic responses have considerably different characteristics than conventional band responses and require acquisition technologies that are capable of meeting far greater requirements. Seismic sources must deliver forces at lower frequencies that are considerably larger than the forces delivered by modern sources at conventional band frequencies in order to achieve comparable signal-to-noise ratios for many traditional interface-related seismic responses. Source efforts that are only comparable to conventional band source efforts are not adequate. Low frequency seismic responses from certain non-interface related impedance changes may be greater, but still require improved low frequency seismic sources.

Automated scoring engines are increasingly being used to score the free-form text responses that students give to questions. Such engines are not designed to appropriately deal with responses that a human reader would find alarming such as those that indicate an intention to self-harm or harm others, responses that allude to drug abuse or sexual abuse or any response that would elicit concern for the student writing the response. Our neural network models have been designed to help identify these anomalous responses from a large collection of typical responses that students give. The responses identified by the neural network can be assessed for urgency, severity, and validity more quickly by a team of reviewers than otherwise possible. Given the anomalous nature of these types of responses, our goal is to maximize the chance of flagging these responses for review given the constraint that only a fixed percentage of responses can viably be assessed by a team of reviewers.

We analyze a within-host model of virus infection with antibody and CD8+ cytotoxic T lymphocyte (CTL) responses proposed by Schwartz et al. (2013). The goal of this work is to gain an overview of the stability of the biologically-relevant equilibria as a function of the model's immune response parameters. We show that the equilibria undergo at most two forward transcritical bifurcations. The model is also explored numerically and results are applied to equine infectious anemia virus infection. In order to arrive at stability of the biologically-relevant endemic equilibrium characterized by coexistence of antibody and CTL responses, the parameters promoting CTL responses need to be boosted over parameters promoting antibody production. This result may seem counter-intuitive (in that a weaker antibody response is better) but can be understood in terms of a balance between CTL and antibody responses that is needed to permit existence of CTLs. In conclusion, an intervention such as a vaccine that is intended to control a persistent viral infection with both immune responses should moderate the antibody response to allow for stimulation of the CTL response.

The unique nonreciprocal responses of superconductors, which stem from the Cooper pairs' quantum condensation, have been attracting attention. Recently, theories of the second-order nonlinear response in noncentrosymmetric superconductors were formulated based on the Bogoliubov-de Gennes theory. In this paper, we study the mechanism and condition for second-order optical responses of time-reversal symmetric superconductors. The numerical results show the characteristic photocurrent and second harmonic generation in the superconducting state. However, the superconductivity-induced nonlinear optical responses disappear under some conditions on pair potential. We show that the coexistence of intraband and interband pairing is necessary for the second-order superconducting optical responses. In addition, the superconducting Berry curvature factor, which is related to a component of Berry curvature in the superconducting state, is essential for the nonlinear responses. Thus, we derived the microscopic conditions where the superconducting nonlinear response appears.

In this paper, we present design, implementation, and effectiveness of generating personalized suggestions for email replies. To personalize email responses based on users style and personality, we model the users persona based on her past responses to emails. This model is added to the language-based model created across users using past responses of the all user emails. A users model captures the typical responses of the user given a particular context. The context includes the email received, recipient of the email, and other external signals such as calendar activities, preferences, etc. The context along with users personality (e.g., extrovert, formal, reserved, etc.) is used to suggest responses. These responses can be a mixture of multiple modes: email replies (textual), audio clips, etc. This helps in making responses mimic the user as much as possible and helps the user to be more productive while retaining her mark in the responses.

Customizing LLMs for a specific task involves separating high-quality responses from lower-quality ones. This skill can be developed using supervised fine-tuning with extensive human preference data. However, obtaining a large volume of expert-annotated data is costly for most tasks. In this paper, we explore a novel method to optimize LLMs using ranking metrics. This method trains the model to prioritize the best responses from a pool of candidates created for a particular task. Rather than a traditional full ordering, we advocate for a partial ordering, as achieving consensus on the perfect order of candidate responses can be challenging. Our partial ordering is more robust, less sensitive to noise, and can be achieved with limited human annotations or through heuristic methods. We test our system's improved response generation ability using benchmark datasets, including textual entailment and multi-document question answering. We conduct ablation studies to understand crucial factors, such as how to gather candidate responses for a specific task, determine their most suitable order, and balance supervised fine-tuning with ranking metrics. Our approach, named Rescue, offers a pro

Categorical responses arise naturally within various scientific disciplines. In many circumstances, there is no predetermined order for the response categories, and the response has to be modeled as nominal. In this study, we regard the order of response categories as part of the statistical model, and show that the true order, when it exists, can be selected using likelihood-based model selection criteria. For predictive purposes, a statistical model with a chosen order may outperform models based on nominal responses, even if a true order does not exist. For multinomial logistic models, widely used for categorical responses, we show the existence of theoretically equivalent orders that cannot be differentiated based on likelihood criteria, and determine the connections between their maximum likelihood estimators. We use simulation studies and a real-data analysis to confirm the need and benefits of choosing the most appropriate order for categorical responses.