Alert button
Picture for Antonis Maronikolakis

Antonis Maronikolakis

Alert button

Politeness Stereotypes and Attack Vectors: Gender Stereotypes in Japanese and Korean Language Models

Jun 16, 2023
Victor Steinborn, Antonis Maronikolakis, Hinrich Schütze

Figure 1 for Politeness Stereotypes and Attack Vectors: Gender Stereotypes in Japanese and Korean Language Models
Figure 2 for Politeness Stereotypes and Attack Vectors: Gender Stereotypes in Japanese and Korean Language Models
Figure 3 for Politeness Stereotypes and Attack Vectors: Gender Stereotypes in Japanese and Korean Language Models
Figure 4 for Politeness Stereotypes and Attack Vectors: Gender Stereotypes in Japanese and Korean Language Models
Viaarxiv icon

Sociocultural knowledge is needed for selection of shots in hate speech detection tasks

Apr 11, 2023
Antonis Maronikolakis, Abdullatif Köksal, Hinrich Schütze

Figure 1 for Sociocultural knowledge is needed for selection of shots in hate speech detection tasks
Figure 2 for Sociocultural knowledge is needed for selection of shots in hate speech detection tasks
Figure 3 for Sociocultural knowledge is needed for selection of shots in hate speech detection tasks
Figure 4 for Sociocultural knowledge is needed for selection of shots in hate speech detection tasks
Viaarxiv icon

This joke is [MASK]: Recognizing Humor and Offense with Prompting

Oct 25, 2022
Junze Li, Mengjie Zhao, Yubo Xie, Antonis Maronikolakis, Pearl Pu, Hinrich Schütze

Figure 1 for This joke is [MASK]: Recognizing Humor and Offense with Prompting
Figure 2 for This joke is [MASK]: Recognizing Humor and Offense with Prompting
Figure 3 for This joke is [MASK]: Recognizing Humor and Offense with Prompting
Figure 4 for This joke is [MASK]: Recognizing Humor and Offense with Prompting
Viaarxiv icon

Analyzing Hate Speech Data along Racial, Gender and Intersectional Axes

May 18, 2022
Antonis Maronikolakis, Philip Baader, Hinrich Schütze

Figure 1 for Analyzing Hate Speech Data along Racial, Gender and Intersectional Axes
Figure 2 for Analyzing Hate Speech Data along Racial, Gender and Intersectional Axes
Figure 3 for Analyzing Hate Speech Data along Racial, Gender and Intersectional Axes
Figure 4 for Analyzing Hate Speech Data along Racial, Gender and Intersectional Axes
Viaarxiv icon

Listening to Affected Communities to Define Extreme Speech: Dataset and Experiments

Mar 22, 2022
Antonis Maronikolakis, Axel Wisiorek, Leah Nann, Haris Jabbar, Sahana Udupa, Hinrich Schuetze

Figure 1 for Listening to Affected Communities to Define Extreme Speech: Dataset and Experiments
Figure 2 for Listening to Affected Communities to Define Extreme Speech: Dataset and Experiments
Figure 3 for Listening to Affected Communities to Define Extreme Speech: Dataset and Experiments
Figure 4 for Listening to Affected Communities to Define Extreme Speech: Dataset and Experiments
Viaarxiv icon

BERT Cannot Align Characters

Sep 20, 2021
Antonis Maronikolakis, Philipp Dufter, Hinrich Schütze

Figure 1 for BERT Cannot Align Characters
Figure 2 for BERT Cannot Align Characters
Figure 3 for BERT Cannot Align Characters
Figure 4 for BERT Cannot Align Characters
Viaarxiv icon

Wine is Not v i n. -- On the Compatibility of Tokenizations Across Languages

Sep 13, 2021
Antonis Maronikolakis, Philipp Dufter, Hinrich Schütze

Figure 1 for Wine is Not v i n. -- On the Compatibility of Tokenizations Across Languages
Figure 2 for Wine is Not v i n. -- On the Compatibility of Tokenizations Across Languages
Figure 3 for Wine is Not v i n. -- On the Compatibility of Tokenizations Across Languages
Figure 4 for Wine is Not v i n. -- On the Compatibility of Tokenizations Across Languages
Viaarxiv icon

Transformers Are Better Than Humans at Identifying Generated Text

Sep 29, 2020
Antonis Maronikolakis, Mark Stevenson, Hinrich Schutze

Figure 1 for Transformers Are Better Than Humans at Identifying Generated Text
Figure 2 for Transformers Are Better Than Humans at Identifying Generated Text
Figure 3 for Transformers Are Better Than Humans at Identifying Generated Text
Figure 4 for Transformers Are Better Than Humans at Identifying Generated Text
Viaarxiv icon