Related papers: How Do Expectations Affect Learning About Fundamen…
We study why overparameterization -- increasing model size well beyond the point of zero training error -- can hurt test error on minority groups despite improving average test error when there are spurious correlations in the data. Through…
Reliable confidence estimation for the predictions is important in many safety-critical applications. However, modern deep neural networks are often overconfident for their incorrect predictions. Recently, many calibration methods have been…
When large language models are aligned via supervised fine-tuning, they may encounter new factual information that was not acquired through pre-training. It is often conjectured that this can teach the model the behavior of hallucinating…
Standard estimators in information provision experiments place more weight on individuals who update their beliefs more in response to new information. This paper shows that, in practice, these individuals who update the most have the…
Humans must flexibly arbitrate between exploring alternatives and exploiting learned strategies, yet they frequently exhibit maladaptive persistence by continuing to execute failing strategies despite accumulating negative evidence. Here we…
Like students facing hard exam questions, large language models sometimes guess when uncertain, producing plausible yet incorrect statements instead of admitting uncertainty. Such "hallucinations" persist even in state-of-the-art systems…
Helping students learn to think like a physicist is an important goal of many introductory physics courses. One characteristic distinguishing more experienced physicists from novice students is that they make better use of problem solving…
We propose an approach to estimate how individuals' expectations influence their responses to a counterfactual change. The approach relies on average partial effects, which recover counterfactual impacts under conditions that we specify. We…
Taking advice from others requires confidence in their competence. This is important for interaction with peers, but also for collaboration with social robots and artificial agents. Nonetheless, we do not always have access to information…
The increasing availability and use of artificial intelligence (AI) tools in educational settings has raised concerns about students' overreliance on these technologies. Overreliance occurs when individuals accept incorrect AI-generated…
Existing learning from demonstration algorithms usually assume access to expert demonstrations. However, this assumption is limiting in many real-world applications since the collected demonstrations may be suboptimal or even consist of…
Past researches show that personality trait is a strong predictor for ones academic performance. Today, mature and verified marker systems for assessing personality traits already exist. However, marker systems-based assessing methods have…
Instruction-tuning enhances the ability of large language models (LLMs) to follow user instructions more accurately, improving usability while reducing harmful outputs. However, this process may increase the model's dependence on user…
Whether examinees' answer changing behavior while taking multiple-choice exams is beneficial or harmful is a long-standing puzzle in the educational and psychological measurement literature. Formalizing the problem using the potential…
When people receive advice while making difficult decisions, they often make better decisions in the moment and also increase their knowledge in the process. However, such incidental learning can only occur when people cognitively engage…
Parameter estimates in misspecified models converge to pseudo-true parameter values, which minimize a population objective function. Pseudo-true values often differ from quantities of economic interest, raising questions of how, if at all,…
The search for a scientific theory of consciousness should result in theories that are falsifiable. However, here we show that falsification is especially problematic for theories of consciousness. We formally describe the standard…
Theoretically as well as experimentally it is investigated how people represent their knowledge in order to make decisions or to share their knowledge with others. Experiment 1 probes into the ways how people 6ather information about the…
Strong student models can learn from weaker teachers: when trained on the predictions of a weaker model, a strong pretrained student can learn to correct the weak model's errors and generalize to examples where the teacher is not confident,…
In this work, we empirically examine human-AI decision-making in the presence of explanations based on predicted outcomes. This type of explanation provides a human decision-maker with expected consequences for each decision alternative at…