CN-122019587-A - Human resource agent prompt word generation method, device and product

Abstract

The application discloses a method, equipment and a product for generating a human resource intelligent agent prompt word. The method comprises the steps of obtaining variable name configuration information, wherein the variable name configuration information comprises at least one variable name, each variable name comprises a main variable name and a sub-variable name, the main variable name is used for identifying a data category, the data category comprises a resume information category and a recruitment information category, the sub-variable name is used for identifying a specific data field under the data category, obtaining a prompt word template, wherein the prompt word template comprises a natural language text and at least one variable defined by a preset identifier, each variable placeholder is defined by the preset identifier, the name of the variable placeholder is consistent with the variable name in the variable name configuration information, analyzing the prompt word template, identifying a plurality of variable placeholders defined by the preset identifier, and extracting the variable name corresponding to each variable placeholder. The embodiment of the application can improve the efficiency of changing the prompt words of the intelligent agent.

Inventors

• HUA XIN
• LIU YUCHENG

Assignees

• 前锦网络信息技术(上海)有限公司

Dates

Publication Date

20260512

Application Date

20251223

Claims (10)

1. 1. The human resource agent prompt word generation method is characterized by comprising the following steps: Obtaining variable name configuration information, wherein the variable name configuration information comprises at least one variable name, each variable name comprises a main variable name and a sub-variable name, the main variable name is used for identifying a data category, the data category comprises a resume information category and a recruitment information category, and the sub-variable name is used for identifying a specific data field under the data category; Obtaining a prompt word template, wherein the prompt word template comprises natural language text and at least one variable defined by a preset identifier, each variable placeholder is defined by the preset identifier, and the name of the variable placeholder is consistent with the variable name in the variable name configuration information; Analyzing the prompt word template, identifying a plurality of variable placeholders defined by the preset identifiers, and extracting variable names corresponding to each variable placeholder; Establishing a mapping relation, and associating each variable name in the variable name configuration information with the variable placeholder identified in the prompt word template to form a one-to-one correspondence between the variable names and the variable placeholders; Acquiring corresponding field values from the user resume and/or recruitment information according to the variable name configuration information; And based on the mapping relation, replacing each variable placeholder in the prompt word template with a corresponding field value to generate an agent prompt word.
2. 2. The method as recited in claim 1, further comprising: And inputting the generated intelligent prompt word into a large model selected by a prompt word writer, so that the selected large model outputs a corresponding result according to the prompt word.
3. 3. The method of claim 1, wherein the primary variable name comprises a resume and a position for indicating that the field value originated from user resume information or recruitment position information, respectively.
4. 4. A method according to claim 3, wherein in the case where the main variable name is a resume, its corresponding sub-variable name includes at least one of gender, age, personal advantage and working year; and in the case that the main variable name is a position, the corresponding sub-variable name comprises at least one of a position name and a position requirement.
5. 5. The method as recited in claim 1, further comprising: And acquiring a prompting word template and corresponding variable name configuration information according to the service codes, wherein the service codes have a one-to-one correspondence with the prompting word template and the corresponding variable name configuration information.
6. 6. The method of claim 1, further comprising parsing the variables in the hint word template, verifying semantic consistency of the variables with variable names based on natural language processing techniques, and if not, issuing a warning or suggesting an adjustment to a hint word writer.
7. 7. The method of claim 5, further comprising dynamically reconstructing a paragraph order or format of the hint word templates based on a type or number of variable names to adapt input requirements or optimize readability of different large models.
8. 8. An electronic device, characterized in that the electronic device is a terminal device or a server, the electronic device comprising a processor and a memory storing computer program instructions, the electronic device implementing the method according to any of claims 1-7 when executing the computer program instructions.
9. 9. A computer program product comprising computer program instructions which, when executed, implement the method of any one of claims 1-7.
10. 10. A computer readable storage medium, characterized in that it stores computer program instructions which, when executed, implement the method of any one of claims 1-7.

Description

Human resource agent prompt word generation method, device and product Technical Field The application relates to the technical field of Internet, and also relates to an artificial intelligence and prompt word generation technology, in particular to a human resource agent prompt word generation method, equipment, a program product and a storage medium. Background At present, the large model and the intelligent agent constructed based on the large model are widely applied to core scenes such as resume screening, person post matching, interview evaluation, occupation planning and the like in the human resource industry. In the implementation process of the applications, the prompt words are used as a core instruction set for guiding the large model to understand tasks and execute special analysis, and the quality and adaptability of the prompt words directly determine the accuracy and practicability of the intelligent agent output. The term templates of the current industry are generally static term templates written by algorithm engineers or business specialists, wherein if specific candidate resume data or post description information is required to be introduced as analysis basis, data fields are usually embedded into the templates in a hard-coded form. When business logic iteration or data analysis dimension changes, for example, an evaluation model is expanded from paying attention to 'working years' to 'gender' and 'age', or a job information field structure is adjusted, a developer is required to intervene to manually modify an application program code for calling a prompt word so as to adapt to new variable input requirements, so that the optimization iteration period of the prompt word is limited by development schedule and cannot respond to business feedback, and the reusability of the same set of prompt word templates in different applications or scenes is poor, management and dispersion are difficult to realize uniform management and control at an enterprise level. Disclosure of Invention In view of the above, embodiments of the present application provide an agent prompt word generating method, an electronic device, a computer readable storage medium and a computer program product based on a prompt word management platform, which solve at least one technical problem. The embodiment of the application provides a human resource intelligent prompt word generation method, which comprises the steps of obtaining variable name configuration information, analyzing the prompt word template, identifying a plurality of variable placeholders defined by the preset identifiers, extracting variable names corresponding to each variable placeholder, establishing a mapping relation, associating each variable name in the variable name configuration information with the variable placeholder identified in the prompt word template to form a variable placeholder value, and generating a corresponding relation between the variable placeholder and the corresponding variable value according to the corresponding relation in the prompt word template or the alternative value of each variable placeholder, wherein the variable name configuration information comprises at least one variable name, each variable name comprises a main variable name and a sub variable name, the main variable name comprises a resume information category and a recruitment information category, the sub variable names are used for identifying specific data fields under the data categories, the prompt word template comprises a natural language text and at least one variable defined by a preset identifier, the name of each variable placeholder is defined by the preset identifier, the name of each variable placeholder is consistent with the variable name in the variable name configuration information, the prompt word template is analyzed, the variable placeholders defined by the preset identifiers are identified, the variable placeholders corresponding to each variable placeholder corresponding to the variable name in the preset identifiers are extracted, the corresponding relation between the variable placeholders in the variable name configuration information and the prompt word template is established, and the corresponding relation between the variable placeholders is formed by the corresponding to the variable placeholders identified in the prompt word values. Optionally, the method according to the embodiment of the application further comprises the step of inputting the generated intelligent prompt word into the large model selected by the prompt word writer, so that the selected large model outputs a corresponding result according to the prompt word. Optionally, according to the method of the embodiment of the present application, the primary variable name includes a resume and a position, which are used to indicate that the field value is derived from user resume information or recruitment position information, respectively. Optionally, according t