Users criticized Gemma 4's weaker performance in Arabic spatial tasks such as understanding tables, accusing prior claims about multilingual models of being misleading.
Based on 1 visible X reactions from 2 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@AdamJaber248 @GoogleDeepMind I've been told multilingual models are worse in Arabic, but not that they are worse at understanding a simple table, or columns or strategize. It is like all I've been told about multilingual models was all lies and coverup, all seeming pretty but no generalization underneath
@AdamJaber248 @GoogleDeepMind I've been told multilingual models are worse in Arabic, but not that they are worse at understanding a simple table, or columns or strategize. It is like all I've been told about multilingual models was all lies and coverup, all seeming pretty but no generalization underneath
Users criticized Gemma 4's weaker performance in Arabic spatial tasks such as understanding tables, accusing prior claims about multilingual models of being misleading.
Based on 1 visible X reactions from 2 accounts; directional sample.
Ask a question below.
Published answers will appear here.