Imprint
Sarah Ball & Phil Hackemann
c/o Institut der Statistik
Ludwig-Maximilians-Universität
Ludwigstraße 33
80539 München
Germany
The operator of this website is responsible for its own content in accordance with applicable laws. However, the operator is not obligated to monitor transmitted or stored third-party information or to investigate circumstances that may indicate illegal activity.
Obligations to remove or block the use of information in accordance with general laws remain unaffected. Liability in this regard is only possible from the point in time at which a specific legal infringement becomes known. Upon becoming aware of such infringements, the content in question will be removed without delay.
Please be aware that parts of this website were created with artificial intelligence.
This website contains links to external third-party websites over whose content the operator has no control. Therefore, no liability is assumed for such external content. The respective provider or operator of the linked websites is solely responsible for their content. The linked pages were reviewed for potential legal violations at the time of linking, and no unlawful content was apparent at that time.
However, continuous monitoring of the content of linked pages is not reasonable without specific indications of a legal violation. Upon becoming aware of any such infringements, the relevant links will be removed promptly.
The content and works published on this website are subject to German copyright law. Any reproduction, processing, distribution, or any form of use beyond the scope permitted by copyright law requires the prior written consent of the respective author or creator. Downloads and copies of this site are permitted for private, non-commercial use only. Only our paper itself (CC-BY) and the video (CC-BY-ND) have been published under a Creative Commons license.
Where the content on this site was not created by the operator, the copyrights of third parties are respected. All names and brands mentioned are the property of their respective owners. Their appearance here is purely for non-commercial academic research, used strictly under fair use guidelines, and does not imply endorsement, affiliation, or sponsorship. Third-party content is identified as such wherever applicable. Should you become aware of a copyright infringement, please inform us accordingly. Upon notification of any such infringement, the relevant content will be removed without delay.
The analyses, evaluations and positions preented on this website and in our paper represent scientific findings and conclusions. This commentary is intended solely for educational, research, and public-interest purposes to foster a critical understanding of alignment methods and AI systems. Nothing on this site is intended to serve as commercial disparagement or a statement of absolute corporate fact outside the scope of the analyzed datasets. Any resemblance to actual persons, living or dead, within the video, which is only a fictional example, is purely coincidental.
The data, responses, and model behaviors documented on this site and in our paper were captured during a specific timeframe by ourselves or by the cited sources. Because the underlying machine learning models and platform algorithms are subject to continuous, unannounced updates by their respective developers, current live platform behaviors may differ from the results published here. All content is provided on an 'as-is' and 'as-available' basis, without warranties of any kind, either express or implied, including but not limited to warranties of accuracy, completeness, reliability, or fitness for a particular purpose.
Under no circumstances shall the authors, or their affiliated academic institutions be held liable for any direct, indirect, incidental, consequential, special, or exemplary damages—including, but not limited to, loss of data, loss of profits, system vulnerabilities, or operational disruptions—arising out of or in connection with the use of, or inability to use, the information, or data presented on this website and in our paper. Users who replicate these research methods or interact with the analyzed AI networks do so entirely at their own risk.