Table 10 Input prompt examples for LLM

Discipline classification

Do not output anything else, just give the tag name, please choose the most appropriate one from the given tag set [‘history’, ‘sports science’, ‘economics’, ‘management’, ‘psychology’, ‘arts’, ‘archaeology’, ‘law’, ‘natural resources and environmental science’, ‘ethnology and cultural studies’, ‘sociology’, ‘statistics’, ‘Marxist theory’, ‘humanistic economic geography’, ‘obscure extinct discipline’, ‘library, information and literature science’, ‘education’, ‘Chinese literature’, ‘foreign literature’, ‘comprehensive social science’, ‘linguistics’, ‘philosophy’, ‘journalism and communication’, ‘political science’, ‘religious studies’, ‘comprehensive university journal’] to classify the following scientific text, only one tag from the given tag set can be assigned to the given text, the given text is: {}

Structural function recognition

Do not output anything else, just give the tag name, please choose the most appropriate tag from the given tag set [‘conclusion’,‘purpose’,‘results’,‘method’,‘limitations’] to classify the following text, the given text is: {}

Named entity recognition

Just output the annotated result, the given tag set is [‘Person’,‘Location’,‘Time’,‘Organization’,‘Metric’,‘Thing’,‘Abstract’], where Person represents names, pronouns, or titles, Location represents place names, Organization represents organization names, Time represents temporal nouns, Metric represents quantifiers other than temporal nouns, Thing represents ordinary objects, and Abstract represents abstracts. Please use these tags to annotate the given text. The output should retain the original text and annotate entities with “<tag>entity</tag>“, for example, if the given text is: “In 1935, a never-before-seen flood of evil water flattened all the buildings on the land of Dongziping to the ground and formed pits, devouring 176 lives.” then you should annotate it as: “<Time>1935</Time>, a never-before-seen flood of <Thing>evil water</Thing> flattened all the buildings on the land of <Thing>Dongziping</Thing> to the ground and formed <Thing>pits</Thing>“. The text you need to process is: {}