S
Sebastian Gehrmann
Researcher at Google
Publications - 87
Citations - 6000
Sebastian Gehrmann is an academic researcher from Google. The author has contributed to research in topics: Computer science & Language model. The author has an hindex of 26, co-authored 63 publications receiving 2233 citations. Previous affiliations of Sebastian Gehrmann include IBM & Bielefeld University.
Papers
More filters
Proceedings ArticleDOI
Causal Analysis of Syntactic Agreement Mechanisms in Neural Language Models
Matthew Finlayson,Aaron Mueller,Sebastian Gehrmann,Stuart M. Shieber,Tal Linzen,Yonatan Belinkov +5 more
TL;DR: This article applied causal mediation analysis to pre-trained neural language models to investigate the magnitude of models' preferences for grammatical inflections, as well as whether neurons process subject-verb agreement similarly across sentences with different syntactic structures.
Proceedings ArticleDOI
Diagnosing AI Explanation Methods with Folk Concepts of Behavior
TL;DR: In this article , a formalism for the conditions of successful explanation of AI is proposed, where success depends not only on what information the explanation contains, but also on what the human explainee understands from it.
Journal ArticleDOI
Performance under pressure in skill tasks: An analysis of professional darts
Marius Ötting,Christian Deutscher,Sandra Schneemann,Roland Langrock,Sebastian Gehrmann,Hendrik Scholten +5 more
TL;DR: Investigating performance under pressure in professional darts as a near-ideal setting with no direct interaction between players and a high number of observations per subject finds no evidence of either choking or excelling under pressure.
Posted Content
Very Highly Skilled Individuals Do Not Choke Under Pressure: Evidence from Professional Darts
Christian Deutscher,Marius Ötting,Roland Langrock,Sebastian Gehrmann,Sandra Schneemann,Hendrik Scholten +5 more
TL;DR: This article investigated performance under pressure in professional darts as a near-ideal setting with no direct interaction between players and a high number of observations per subject and found strong evidence for an overall improved performance for nearly all players in the sample.
Journal ArticleDOI
Evaluating an Automated Mediator for Joint Narratives in a Conflict Situation
Massimo Zancanaro,Oliviero Stock,Gianluca Schiavo,Alessandro Cappelletti,Sebastian Gehrmann,Daphna Canetti,Ohad Shaked,Shani Fachter,Rachel Yifat,Ravit Mimran,Patrice L.,Weiss +11 more
TL;DR: In this article, an automated cognitive tutor is used to support remote interactions between pairs of participants as they contribute to a shared story in their own language, and the tutor guides the participants through a controlled escalation/de-escalation process during the development of a joint narrative.