52% Yes, a signiicant crisis 3% No, there is no crisis 7% Don't know 38% Yes, a slight crisis 38% Yes, a slight crisis 1,576 RESEARCHERS SURVEYED M ore than 70% of researchers have tried and failed to reproduce another scientist's experiments, and more than half have failed to reproduce their own experiments. Those are some of the telling figures that emerged from Nature's survey of 1,576 researchers who took a brief online questionnaire on reproducibility in research. The data reveal sometimes-contradictory attitudes towards reproduc-ibility. Although 52% of those surveyed agree that there is a significant 'crisis' of reproducibility, less than 31% think that failure to reproduce published results means that the result is probably wrong, and most say that they still trust the published literature. Data on how much of the scientific literature is reproducible are rare and generally bleak. The best-known analyses, from psychology 1 and cancer biology 2 , found rates of around 40% and 10%, respectively. Our survey respondents were more optimistic: 73% said that they think that at least half of the papers in their field can be trusted, with physicists and chemists generally showing the most confidence. The results capture a confusing snapshot of attitudes around these issues, says Arturo Casadevall, a microbiologist at the Johns Hopkins Bloomberg School of Public Health in Baltimore, Maryland. "At the current time there is no consensus on what reproducibility is or should be. " But just recognizing that is a step forward, he says. "The next step may be identifying what is the problem and to get a consensus. "

The aim of the study was to describe and compare the school participation and social networks of children with physical disabilities and complex communication needs (Group CCN), children with physical disabilities only (Group PD), and children with typical development (Group TD). The 39 participants, 10-15 years of age, were observed for 4 hours at school. School staff and the parent and/or child provided information on children's social networks. A striking observation was that, while participants in Group TD continuously conversed and socialized with peers inside and outside classrooms; those in Group CCN rarely used aided AAC, were provided with limited communication opportunities at school, and had fewer acquaintances and friends. Findings warrant intervention at the participation level at school and in the community.

Context:The use of levothyroxine to reduce thyroid size in pediatric patients with goiter due to chronic autoimmune thyroiditis (AIT) remains controversial. In overtly hypothyroid patients, reductions in thyroid volume have been reported, whereas the effect in subclinically hypothyroid and euthyroid patients is less clear. Objective:The objective of the study was to evaluate the effect of levothyroxine treatment on thyroid size (determined with thyroid ultrasonography) in children and adolescents with AIT. Design and Setting: This study included patients with AIT treated at a university hospital outpatient clinic between 1987 and 2004.Patients: Ninety children with AIT (73 girls and 17 boys, aged 6.1-17.7 yr) were included in the study.Intervention: Intervention was treatment with levothyroxine for a median 2.8 yr (range 0.5-10.2).Main Outcome Measure: Change in thyroid volume SD score (SDS) during the study period was measured.Results: Median thyroid volume SDS was reduced in patients euthyroid (Ϫ0.4 SDS, P Ͻ 0.001), subclinically hypothyroid (Ϫ1.4 SDS, P Ͻ 0.001), and overtly hypothyroid (Ϫ1.8 SDS, P Ͻ 0.002) at diagnosis of AIT. Both hypothyroid and euthyroid patients with goiter (thyroid volume Ͼ 2.0 SDS) at baseline reduced their median thyroid volume SDS (Ϫ1.6 and Ϫ0.9, respectively, P Ͻ 0.001). Hypothyroid patients without goiter also reduced median thyroid volume SDS (Ϫ1.2, P Ͻ 0.004), whereas no change was noticed in euthyroid children without goiter. Conclusions:Levothyroxine treatment is effective in reducing thyroid volume in pediatric patients and is suggested in treatment of goiter caused by AIT, especially in cases of hypothyroid, but also in euthyroid children.

Given the broad capabilities of large language models, it should be possible to work towards a general-purpose, text-based assistant that is aligned with human values, meaning that it is helpful, honest, and harmless. As an initial foray in this direction we study simple baseline techniques and evaluations, such as prompting. We find that the benefits from modest interventions increase with model size, generalize to a variety of alignment evaluations, and do not compromise the performance of large models. Next we investigate scaling trends for several training objectives relevant to alignment, comparing imitation learning, binary discrimination, and ranked preference modeling. We find that ranked preference modeling performs much better than imitation learning, and often scales more favorably with model size. In contrast, binary discrimination typically performs and scales very similarly to imitation learning. Finally we study a 'preference model pre-training' stage of training, with the goal of improving sample efficiency when finetuning on human preferences.

No abstract