Human Performance on Chinese Spatial Semantic Judgment

Anonymous

Abstract: With the emergence of ChatGPT, large-scale language models seem to possess cognitive abilities similar to humans. This paper mainly focuses on the comparative analysis of human-machine testing on the task of judging the correctness and incorrectness of Chinese spatial semantics, including the process of human testing, the source, scale, and text characteristics of human testing data, and the comparison of human-machine testing accuracy, etc. By summarizing the typical features presented by human-machine on spatial semantic topics, this paper tries to analyze whether the machine has human-like spatial language understanding ability.

Paper Type: long

Research Area: Resources and Evaluation

Contribution Types: Data resources, Data analysis

Languages Studied: English, Chinese

Consent To Share Submission Details: On behalf of all authors, we agree to the terms above to share our submission details.