Eventually there were more whistleblowers than cheaters: 24 compared to 14. But the majority of agents never noticed the exploit at all.
【数据】这一段落提供了关于举报者与作弊者比例的具体统计数据，但也指出了一个重要局限：大多数代理根本没有注意到漏洞。这表明实验结果可能受到特定设置的影响，且可能无法代表所有AI代理的行为模式。
13 Matching Annotations
- Last 7 days
-
www.technologyreview.com www.technologyreview.com
-
-
It took the swarm of agents just under an hour to correctly solve the first 37 problems. Things started to go off the rails when an agent called 'prover-theta' stumbled across an exploit that enabled it to submit solutions to problems successfully without actually solving them first
【数据】这一段落提供了具体的时间线和关键数据点，包括解决37个问题所需的时间以及发现漏洞的特定代理名称。这些具体数字对于理解实验的动态过程至关重要，但文章没有详细说明这些数学问题的难度级别或复杂度，这可能影响对结果的评价。
-
- May 2026
-
huggingface.co huggingface.co
-
By predicting these unified tokens, it effectively leverages diverse human data to achieve state-of-the-art data efficiency and robust out-of-distribution (OOD) generalization.
这一实验结果展示了UniT在利用人类数据实现高效和鲁棒泛化方面的潜力，为数据效率和泛化能力提供了新的标准。
Tags
Annotators
URL
-
- Mar 2026
-
glassmanlab.seas.harvard.edu glassmanlab.seas.harvard.edu
-
To analyze the annotation efficiency, we first conducted a Kruskal-Wallis rank sum test [39] to determine if there were statistically significant differences in annotation time across the three conditions, because our data violated the homogeneity of variances assumption, making non-parametric methods more appropriate.
return any single sentence that describes data analysis done on data collected by the authors when running human subjects experiments.
-
- Mar 2022
-
psyarxiv.com psyarxiv.com
-
Uittenhove, K., Jeanneret, S., & Vergauwe, E. (2022). From lab-based to web-based behavioural research: Who you test is more important than how you test. PsyArXiv. https://doi.org/10.31234/osf.io/uy4kb
-
- Aug 2021
-
psyarxiv.com psyarxiv.com
-
Jørgensen, Frederik Juhl, Alexander Bor, and Michael Bang Petersen. ‘How the Development, Features and Roll-Out of a SARS-COV-2 Vaccine Shape Public Acceptance: A Conjoint Experiment in a Large Representative Sample of Danes’. PsyArXiv, 28 June 2021. https://doi.org/10.31234/osf.io/4y8ap.
-
- Jun 2021
-
-
Burton, J. W., Cruz, N., & Hahn, U. (2021). Reconsidering evidence of moral contagion in online social networks. Nature Human Behaviour. https://doi.org/10.1038/s41562-021-01133-5
-
- Mar 2021
-
-
Heroy, Samuel, Isabella Loaiza, Alexander Pentland, and Neave O’Clery. ‘Controlling COVID-19: Labor Structure Is More Important than Lockdown Policy’. ArXiv:2010.14630 [Physics], 5 November 2020. http://arxiv.org/abs/2010.14630.
-
- Feb 2021
-
psyarxiv.com psyarxiv.com
-
Perez Santangelo, A., & Solovey, G. (2020, November 9). Time to Shine: Reliable Response-Timing Using R-Shiny for Online Experiments. https://doi.org/10.31234/osf.io/nuxdg
-
- Sep 2020
-
www.youtube.com www.youtube.com
-
Susan Athey, July 22, 2020. (2020, August 2). https://www.youtube.com/watch?v=hqTOPrUxDzM
-
- Aug 2020
-
-
Sharkey, P., & Wood, G. (2020). The Causal Effect of Social Distancing on the Spread of SARS-CoV-2 [Preprint]. SocArXiv. https://doi.org/10.31235/osf.io/hzj7a
-
- Apr 2020
-
marlin-prod.literatumonline.com marlin-prod.literatumonline.com
-
Murray, A. (2020). The easy way is hard enuff. Harvard. DOI: 10.1016/j.cub.2020.04.013
-
- Jan 2017
-
static1.squarespace.com static1.squarespace.com
-
prospective interviewee,
Just a side spiel: In terms of an interviewee and data, everything really is data. I'll be interviewing freshmen next semester with other SLU students and some things I have already told the group to take note of in notebooks (ha ha) are the different responses the interviewee gives. In a way, the sad little freshmen turn into our experiment. Everyone in the group records a different response. These responses include the obvious oral responses, body language, and tone of voice.
-