WO2025026189A1 - Method and apparatus for prompt word management, device, and storage medium - Google Patents

Abstract

Embodiments of the present disclosure provide a method and apparatus for prompt word management, a device, and a storage medium. The method comprises: in response to receiving a prompt word editing request, presenting an editing interface, wherein the editing interface at least comprises an input area for receiving a target prompt word; and in response to receiving a prompt word publishing request, encapsulating the target prompt word into a callable target task at least on the basis of a user input in the editing interface, wherein the calling of the target task triggers the target prompt word to be inputted to a model. Thus, prompt words can be flexibly designed and managed, and the stability, reliability, convenience and high efficiency of prompt word management and model calling in practical application can be improved.

Description

Method, device, equipment and storage medium for prompt word management

This application claims priority to the Chinese invention patent application entitled “Methods, devices, equipment and storage media for prompt word management” and application number 202310945515.8, filed on July 28, 2023. The entire contents of this application are incorporated by reference into this application.

Technical Field

Example embodiments of the present disclosure generally relate to the field of computers, and more particularly, to methods, devices, apparatuses, and computer-readable storage media for prompt word management.

Background Art

With the development of machine learning technology, machine learning models can be used to perform tasks in a variety of application environments. Since the development and maintenance of high-performance models requires a lot of technical and financial investment, including huge data sets, high-performance computing resources, and professional research teams, Model as a Service (MaaS) has gradually become a feasible solution. Users can input prompt words into the model in the cloud to obtain the model output for this prompt word. Therefore, the accuracy of the prompt word will affect the output of the model. How to manage the prompt word conveniently and quickly is a technical problem to be explored at present.

Summary of the invention

In a first aspect of the present disclosure, a method for prompt word management is provided. The method comprises: in response to receiving a prompt word editing request, presenting an editing interface, wherein the editing interface at least comprises an input area for receiving a target prompt word; and in response to receiving a prompt word publishing request, encapsulating the target prompt word into a callable target task based at least on user input in the editing interface, wherein the call of the target task triggers the target prompt word to be input into a model.

In a second aspect of the present disclosure, a device for prompt word management is provided. The device includes: an interface presentation module, configured to present an editing interface in response to receiving a prompt word editing request, wherein the editing interface at least includes an input area for receiving a target prompt word; and a publishing module, configured to encapsulate the target prompt word into a callable target task in response to receiving a prompt word publishing request, at least based on user input in the editing interface, wherein the call of the target task will trigger the target prompt word to be input into the model.

In a third aspect of the present disclosure, an electronic device is provided. The device includes at least one processing unit; and at least one memory, the at least one memory is coupled to the at least one processing unit and stores instructions for execution by the at least one processing unit. When the instructions are executed by the at least one processing unit, the electronic device executes the method of the first aspect.

In a fourth aspect of the present disclosure, a computer-readable storage medium is provided, wherein a computer program is stored on the medium, and the computer program can be executed by a processor to implement the method of the first aspect.

It should be understood that the content described in this section is not intended to limit the key features or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description.

BRIEF DESCRIPTION OF THE DRAWINGS

The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, wherein:

FIG1 is a schematic diagram showing an example environment in which embodiments of the present disclosure can be implemented;

2A to 2J are schematic diagrams showing example interfaces of a prompt word management system according to some embodiments of the present disclosure;

FIG3 is a schematic diagram showing a process of prompt word management according to some embodiments of the present disclosure;

FIG4 shows a flowchart of a process for prompt word management according to some embodiments of the present disclosure;

FIG5 shows a schematic structural block diagram of an apparatus for prompt word management according to certain embodiments of the present disclosure; and

FIG. 6 illustrates an electronic device in which one or more embodiments of the present disclosure may be implemented.

DETAILED DESCRIPTION

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments set forth herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

In the description of the embodiments of the present disclosure, the term "including" and similar terms should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The term "some embodiments" should be understood as "at least some embodiments". Other explicit and implicit definitions may also be included below.

Herein, unless explicitly stated, executing a step “in response to A” does not mean executing the step immediately after “A” but may include one or more intermediate steps.

It is understandable that the data involved in this technical solution (including but not limited to the data itself, the acquisition, use, storage or deletion of the data) shall comply with the requirements of relevant laws, regulations and relevant provisions.

It is understandable that before using the technical solutions disclosed in the various embodiments of the present disclosure, the types, scopes of use, usage scenarios, etc. of the information involved in the present disclosure should be informed to relevant users and their authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations. The relevant users may include any type of right holders, such as individuals, enterprises, and groups.

For example, in response to receiving an active request from a user, a prompt message is sent to the relevant user to clearly prompt the relevant user that the operation requested to be performed will require obtaining and using the information of the relevant user, so that the relevant user can independently choose whether to provide information to software or hardware such as an electronic device, application, server or storage medium that executes the operation of the technical solution of the present disclosure based on the prompt message.

As an optional but non-limiting implementation, in response to receiving an active request from a relevant user, a prompt message is sent to the relevant user, for example, in the form of a pop-up window, in which the prompt message may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to select "agree" or "disagree" to provide information to the electronic device.

It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet relevant laws and regulations may also be applied to the implementation of the present disclosure.

FIG1 shows a schematic diagram of an example environment 100 in which embodiments of the present disclosure can be implemented. In the example environment 100, a prompt word management system 120 is running in an electronic device 110. A user 140 can interact with the prompt word management system 120 via the electronic device 110 and/or an attached device of the electronic device 110. The prompt word management system 120 can provide integration of various types of functions or components to facilitate people to create, modify, adjust, and other activities of prompt words.

In some embodiments, the prompt word management system 120 can be downloaded and installed as an application on the electronic device 110. In some embodiments, the prompt word management system 120 can also be accessed through other methods, such as through a web page.

In the environment 100 of FIG1 , if the prompt word management system 120 is started, the electronic device 110 may present an interface 150 of the prompt word management system 120. The interface 150 may include various user interfaces that the prompt word management system 120 may provide. The interface 150 is sometimes also referred to as a client interface.

In some embodiments, the prompt word management system 120 can realize the generation, verification, evaluation, etc. of prompt words with the help of the model 130. In this article, the prompt word (prompt) is the input of the model or the input template, which can be natural language and/or other modal data, used to guide or stimulate the model to complete specific processing to obtain the desired model output. The prompt word management system 120 can interact with the model 130 to provide model input to the model 130 and obtain the corresponding model output from the model 130. The model 130 can run locally on the electronic device 110 or on a remote server. In some embodiments, the model 130 can be a machine learning model, a deep learning model, a learning model, a neural network, etc. In some embodiments, the model 130 can be based on a language model (LM). The language model can have question-answering capabilities by learning from a large amount of corpus. The model 130 can also be based on other appropriate models.

The electronic device 110 can be any type of device with computing capabilities, including terminal devices and/or server devices. The terminal device can be any type of mobile terminal, fixed terminal or portable terminal, including mobile phones, desktop computers, laptop computers, notebook computers, netbook computers, tablet computers, media computers, multimedia tablets, personal communication systems (PCS) devices, personal navigation devices, personal digital assistants (PDAs), audio/video players, digital cameras/camcorders, positioning devices, television receivers, radio broadcast receivers, e-book devices, gaming devices, or any combination of the foregoing, including accessories and peripherals of these devices or any combination thereof. The server device can include, for example, a computing system/server, such as a mainframe, an edge computing node, a computing device in a cloud environment, and the like. In some embodiments, the prompt word management system 120 can be implemented based on a cloud service.

It should be understood that the structure and function of the various elements in the environment 100 are described for exemplary purposes only and do not imply any limitation on the scope of the present disclosure.

The core concept of Model as a Service is to deploy models, especially excellent head models, in the cloud. Users can call models through APIs or SDKs to implement various functions. R&D costs can be reduced, and advanced AI technologies can be quickly deployed and applied. Compared with the traditional method of training different models for tasks, the advantages of Model as a Service include lower cost, more flexibility (users can choose appropriate models and services according to their needs and adjust usage plans at any time), easy integration (Model as a Service can be easily integrated with existing systems through APIs or SDKs, etc., without large-scale transformation), continuous updates (cloud models are regularly updated and optimized to ensure that users can enjoy the latest technological achievements), cross-platform compatibility (support for multiple platforms and devices, facilitating users to carry out cross-platform development), security and privacy protection, etc.

In addition, through the use of platforms and/or systems, the call of the model becomes a very simple way. At present, the use of models is mostly based on instructions and question-and-answer interactions. For users, they only need to design a prompt word to provide to the model to generate answers. However, since the design of prompt words needs to fully consider a lot of information, such as contextual information, domain knowledge, and the inherent limitations of the model, and the design of prompt words needs to be continuously iterated and optimized. Therefore, in practical applications, designing a high-quality prompt word to guide the model to generate more accurate and relevant answers is a very challenging task.

According to an embodiment of the present disclosure, an improved scheme for prompt word management is proposed. In the embodiment of the present disclosure, in response to receiving a prompt word editing request, an editing interface that at least includes an input area for receiving a target prompt word is presented. In response to receiving a prompt word publishing request, the target prompt word is encapsulated into a callable target task based at least on user input in the editing interface. The call to the target task will trigger the target prompt word to be input into the model. Thus, the prompt words can be flexibly designed and managed, and the prompt words can be encapsulated into a callable task, which makes the call of the model very simple, and helps to improve the stability, reliability, convenience and efficiency of prompt word management and model call in practical applications.

Some example embodiments of the present disclosure will be described in detail below with reference to examples of the accompanying drawings.

In an embodiment of the present disclosure, when the prompt word management system 120 is in operation, the prompt word management system 120 presents an editing interface including at least an input area for receiving a target prompt word in response to receiving a prompt word editing request. The prompt word editing request here may include, for example, but is not limited to, a task creation request for a new prompt word and a task modification request for an existing prompt word. Taking the prompt word editing request as a creation request for a new prompt word as an example, the target prompt word is the new prompt word to be created, and the prompt word management system 120 may present an editing interface including an input area for receiving the new prompt word.

Further, in response to receiving the prompt word publishing request, the prompt word management system 120 obtains the callable target task based on the target prompt word at least based on the user input in the editing interface. Exemplarily, the target prompt word can be encapsulated into a section of code corresponding to the callable task, and by triggering the call to this section of code, the corresponding target prompt word can be input to the model. Exemplarily, if the target prompt word indicates the prompt word associated with the article translation, the target task indicating the translation of an article can be generated based on the call parameters of the target prompt word and the model, and the model then performs the target task. About the specific mode of the model to perform the target task, in response to calling the target task, the target prompt word is input to the model. The model can perform the operation indicated by the target task based on the target prompt word. For example, based on the target prompt word, the translation of an article is performed.

The following will describe example interfaces of the prompt word management system of some embodiments of the present disclosure in conjunction with Figures 2A to 2J. Figures 2A to 2J are schematic diagrams of examples of editing interfaces according to some embodiments of the present disclosure. The editing interfaces shown in Figures 2A to 2J may correspond to the prompt word management system 120, which integrates multiple different types of components. It should be understood that the editing interfaces shown in Figures 2A to 2J are only example interfaces, and various interface designs may actually exist. The various graphical elements in the interface may have different arrangements and different The same visual representation may be used, one or more elements may be omitted or replaced, and one or more other elements may also exist. The embodiments of the present disclosure are not limited in this respect.

In some embodiments, when the prompt word management system 120 is in operation, the prompt word management system 120 may present an editing interface 200A for receiving a prompt word editing request as shown in FIG2A. In some embodiments, since the prompt word management system 120 supports multi-environment applications, before using the prompt word management system to perform tasks such as creating tasks for new prompt words, it is necessary to first determine the target environment for the task application. In some embodiments, in response to detecting a trigger operation on the environment selection control 201, a plurality of candidate environments to be selected are presented (area A online, area B online, and area C online as shown in FIG2A). The target environment for the task application may be determined in response to the selection operation on the candidate environment. For example, in response to receiving a click operation on the candidate environment on the area A line, it is determined as the target environment for the task application. Thus, a target environment that meets the regional requirements may be selected based on user needs and the actual application environment.

In some embodiments, a task list including at least one created task can also be presented on the prompt word management system 120. As shown in FIG. 2A, two task lists, "My Task List" and "Public Task List", are presented in the editing interface 200A. In one example, "My Task List" can be, for example, a task list including at least one task created by the current user corresponding to the prompt word management system 120, and "Public Task List" can be, for example, a task list including at least one task created by other users other than the current user and disclosed so that the current user can obtain it. It can be understood that other task lists can also be presented, or only one task list containing all tasks can be presented, and the present disclosure does not limit this. In response to detecting a trigger operation on "My Task List" or on "Public Task List", it can be switched to present at least one task in the corresponding task list. An example of presenting at least one task included in "My Task List" is shown in FIG. 2A, and "My Task List" includes Task A, Task B, Task C and Task D.

The task list includes a task ID and a task name that can be used to identify the task, a task keyword for indicating important task information (such as a required model, task purpose, etc.), the creation time, update time, and creator of the task. The creation time of the task is the time when the task was initially created, and the update time of the task is the time when the task was last updated. In some embodiments, at least one operation control can be presented in association with at least one task included in the task list. In some embodiments, the editing interface 200A includes at least one delete control 205 indicating the deletion of the task, and the at least one delete control 205 corresponds to at least one task in the task list. In response to detecting a trigger operation on the delete control 205, it can be determined that a task deletion request for the corresponding task has been received. In response to the task deletion request, the corresponding task will be removed from the editing interface 200A.

In some embodiments, the editing interface 200A may further include at least one modification control 204 associated with at least one task included in the task list, and the at least one modification control 204 corresponds to at least one task in the task list. In some embodiments, the editing interface 200A also includes a creation control 203. It is understood that the editing interface 200A may also include other operation controls, which are not limited by the present disclosure. In some embodiments, in response to detecting a trigger operation to the creation control 203, it may be determined that a task creation request for a new prompt word is received. In response to detecting a trigger operation to the modification control 204, it may be determined that a task modification request for an existing prompt word is received.

In some embodiments, since the task is associated with the prompt word, the task creation request and/or task modification request for the task can be understood as a creation request and/or modification request for the prompt word. Such a creation request and modification request are both prompt word editing requests for the prompt word. That is, in response to detecting the creation control 203 and/or modification request, The triggering operation of the control 204 can receive a prompt word editing request (creation request and/or modification request) for the prompt word.

In some embodiments, the prompt word editing request can also be received in other ways. Exemplarily, it can be determined that the prompt word editing request is received in response to detecting a voice or gesture corresponding to the prompt word editing request, or detecting a click operation on a corresponding physical control. The voice and gesture here can be pre-set, or include pre-set keywords/key actions. For example, in response to detecting a voice "create task" or "modify task A", etc., it is determined that the prompt word editing request is received.

It should be noted that, in the case where the prompt word editing request is a modification request for a created task (for example, a task modification request received in response to detecting a trigger operation on the modification control 204), the prompt word management system 120 can also determine the authority of the current user. Specifically, a task can be created by the current user (at least one task included in the "My Task List" as shown in FIG. 2A), or it can be created by other users and then made public so that the current user can obtain it (at least one task included in the "Public Task List" as shown in FIG. 2A). For tasks created by the user himself, it can be understood that the user has the highest authority to edit it, and the user can modify it. However, for tasks created by other users and made public so that the current user can obtain it, the prompt word management system 120 needs to determine the authority of the current user to determine whether the current user has the qualifications to edit it. It can be understood that different users may have different authorities for the same task, for example, user A has editing authority, while user B only has use authority but not editing authority. For created tasks, the prompt word management system 120 can present the editing interface only when it is determined that the user has editing authority for the created task. It is understood that, in the case of creating a new task, the prompt word management system 120 can directly present an editing interface without verifying the editing authority. Thus, different users can have different permissions for the same task, which helps different users in the team to share tasks, achieve rapid collaboration and effective communication, and help improve the efficiency of prompt word management.

In some embodiments, the editing interface 200A may further include a help control 202. In response to detecting a trigger operation on the help document help control 202, the interface 200A may switch to a document presentation interface (not shown in the figure), which may include text information and/or image information for prompting the user about the interaction method of the prompt word management system 120, the prompt word creation operation steps, the prompt word modification operation steps, etc. The trigger operation here includes but is not limited to a click operation, a long press operation, a slide operation, a double-click operation, etc.

In some embodiments, after receiving the prompt word edit request, in response to the prompt word edit request, an editing interface 200B as shown in Figure 2B can be presented. Exemplary, in response to detecting a trigger operation for creating a control 203 and/or modifying a control 204, an editing interface 200B as shown in Figure 2B can be presented. The editing interface 200B may include a task definition area 210. In some embodiments, the task definition area 210 is used to receive a task identifier of a target task, and the task identifier may be used for subsequent calls to the target task. Specifically, when the target task needs to be called, the call to the target task may be performed based on the task identifier of the target task, which helps to improve the efficiency of task calls. The task identifier here includes at least one of a task name and a task description. It is understandable that the task identifiers of different tasks are different (that is, at least one of a task name and a task description is different).

The editing interface 200B includes an input area 215. In some embodiments, the input area 215 can be used to receive a template of a target prompt word, and the template can be used to define the structure of the target prompt word, including one or more variables contained in the target prompt word. In this way, compared with defining a one-time, complete prompt word content, a template can be designed to allow the values of corresponding variables to be filled into the template in a specific application to obtain the input to the target prompt word. The prompt word management system 120 detects the user input in the input area 215 and determines the user input as the template of the received target prompt word. The input area 215 can also present the template of the received target prompt word to the user. As shown in FIG2B , the template of the target prompt word received in the input area 215 is, for example, "It is necessary to implement the echo function according to the message, and the original text must be returned without adding any other content. Message: {{msg}}echo:".

Since the target prompt word may contain at least one variable (message as shown in FIG. 2B), in order to define the variable so that the subsequent assignment of the variable is more accurate, the editing interface 200B may also include a variable definition area 220. In some embodiments, the variable definition area 220 is used to receive the definition of at least one variable involved in the target prompt word, including but not limited to the name of the variable (i.e., the variable name), the description of the variable, the variable default value, etc. As shown in FIG. 2B, the definition of the variable may indicate that the variable name is "msg", the variable description is "message returned by echo", and if the user input is received in the "variable default value" column, the received user input may be determined as the default assignment of the variable. If the user input is not received in the "variable default value", after the prompt word is released, in response to detecting the call to the target task, an assignment prompt information for prompting the user to assign a value to at least one variable may be presented. The presentation method here includes but is not limited to voice, vibration, message box, pop-up window, etc. The user can manually assign a value to at least one variable based on the assignment prompt information, or instruct the prompt word management system 120 to assign a value based on the associated database.

The editing interface 200B also includes a model configuration area 230. In some embodiments, the model configuration area 230 can be used to receive a selection of a target model. When the target task is subsequently called, the target keyword will be input into the target model selected in the model configuration area 230. As shown in Figure 2B, the model configuration area 230 may include a model selection entry 231. In response to detecting a click operation on the model selection entry 231, multiple models can be presented, and then in response to a selection operation on a certain model (such as model A), it is determined as the target model to which the corresponding prompt word will be input. In some embodiments, the model selection entry 231 can also present a pre-set default model, which is the target model.

In some embodiments, in order to ensure the accuracy of subsequent task calls and task execution, the model configuration area 230 is also used to receive the definition of processing parameters for the target model. As shown in Figure 2B, the model configuration area 230 can include setting entries for parameters such as degree, maximum length, stop sequence, top, existence penalty, frequency penalty, etc.

Exemplarily, the parameters that can be configured in the model configuration area 230 may include a temperature parameter. The temperature can be used to control the diversity and creativity of the generated text. Its value range is usually between 0.1 and 1, indicating the size of randomness when generating text. When the value is low, the generated text will be more conservative and predictable, and vice versa, it will be more bold and unpredictable. For example, if the value of the temperature is set to 0.5, the generated text will have certain creativity and changes, but still maintain certain coherence and readability. And if the value of the temperature is set to 1, the generated text will be more random and unpredictable, and may contain some strange sentences and meaningless content. Therefore, it is necessary to adjust the value of the temperature according to specific needs and application scenarios to achieve the best effect.

For example, the parameters that can be configured in the model configuration area 230 may also include a maximum length parameter. The maximum length refers to the maximum length of the generated text. The maximum length of the generated text can be set to control the length of the generated text. If the maximum length is set to 100, the generated text will not exceed 100 characters. This parameter can help control the length of the generated text to adapt to different application scenarios. For example, if you need to generate some short text, you can set the maximum length to 50, and if you need to generate some long text, you can set the maximum length to 500 or higher.

Exemplarily, the parameters that can be configured in the model configuration area 230 may also include a stop sequence parameter. The stop sequence is used to specify a sign to stop when generating text. When any of the signs specified in the stop sequence appears in the generated text, the API stops generating text and returns the result. This parameter can be used to control the length and content of the generated text, as well as to avoid generating meaningless or incoherent text. For example, if the stop sequence is set to common period signs such as ".", "?", "!", the API will stop after generating a complete sentence to avoid generating too long or incomplete text.

Exemplarily, the parameters that can be configured in the model configuration area 230 may also include a Top P parameter. Top is used to control the diversity and predictability of the generated text. Specifically, Top is a decimal between 0 and 1, indicating that only the top P words with the highest probability are considered when generating text. For example, if Top is set to 0.5, only 50% of the words with the highest probability are considered when generating text, while words with lower probabilities are ignored. This can make the generated text more predictable, but it may also cause the text to be too monotonous and lack creativity. Therefore, according to the specific application scenarios and requirements, the value of Top can be appropriately adjusted to balance the diversity and predictability of the text.

Exemplarily, the parameters that can be configured in the model configuration area 230 may also include a presence penalty parameter. The presence penalty is used to control whether specific words or phrases are included in the generated text. The higher the value of the presence penalty, the less likely it is that these words or phrases will be included in the generated text. This parameter can be used to control the style and content of the generated text to make it more in line with specific needs. For example, if you need to generate text about healthy eating, you can set a presence penalty to ensure that unhealthy food or drinks do not appear in the generated text.

Exemplarily, the parameters that can be configured in the model configuration area 230 may also include a frequency penalty parameter. The frequency penalty is used to control the frequency of repeated words in the generated text. If a higher frequency penalty value is set, the generated text will try to avoid repeating the same words, making the generated text more diverse and rich. On the contrary, if a lower frequency penalty value is set, a large number of repeated words may appear in the generated text, making the generated text simpler and more repetitive. Therefore, by adjusting the frequency penalty parameter, the diversity and complexity of the generated text can be controlled.

Can be in response to the setting operation received at least one setting entrance, determine the processing parameters of target model about degree, maximum length, stop sequence, top, existence penalty, frequency penalty etc.In certain embodiments, default parameter can be provided, and the user can modify the default parameter to determine the processing parameters of target model.It can be understood that for different models, the processing parameters that can be configured can be different.In certain embodiments, when the target task is subsequently called, the target model will process the target prompt word to perform the target task according to the defined processing parameters.In certain embodiments, the target model determined in the model configuration area 230 model configuration area 230 can also be used for the subsequent verification, evaluation etc. of the template of the target prompt word.

As shown in Fig. 2B, in some embodiments, the editing interface 200B may also include an editing control presentation area 240. A save control 241 may be presented in the editing control presentation area 240. In some embodiments, in response to detecting a trigger operation to the save control 241, it may be determined that a save request to the target prompt word is received. Information such as the template of the target prompt word determined in the editing interface 200B, the definition of the variables, the processing parameters of the model, etc. may be saved. In some embodiments, this information will be saved in a draft box so that editing can be continued next time.

A validation control 242 may also be presented in the edit control presentation area 240. In some embodiments, before the target prompt word is published, that is, before receiving the prompt word publishing request, in response to detecting the validation control 242 In response to the trigger operation of the verification control 242, it can be determined that a verification request for the target prompt word is received. In response to receiving the verification request for the target prompt word, at least based on the user input in the editing interface, the target prompt word is input into the model, and the model output for the target prompt word is received from the model. Further, the received model output can also be presented in the editing interface 200B. As shown in Figure 2C, in response to receiving the trigger operation of the verification control 242, it can be determined that a verification request for the target prompt word is received. Further, in response to the verification request for the target prompt word, a verification area 250 is presented in the editing interface 200B. The verification area 250 includes a variable presentation area 251, and the variable presentation area 251 can present the assignment of the variables in the template of the target prompt word based on the user's input (this assignment is the user's input). The verification area 250 also includes a template presentation area 252 for presenting the template of the target prompt word and an output presentation area 253 for presenting the model output for the target prompt word received from the model. In some embodiments, the verification area 250 also presents a cancel control 254 and a start debugging control 255. In response to detecting the trigger operation of cancel control 254, the verification of the target prompt word can be ended. It can be understood that after the verification is canceled, the editing interface 200B will stop presenting the verification area 250. In response to detecting the trigger operation of the start debugging control 255, the verification of the target prompt word can be started, and the assignment of the variable presented in the variable presentation area 251 is assigned to the template shown in the template presentation area 252 to obtain the target prompt word, and this target prompt word will be input into the model. The model output for the target prompt word can be received from the model and presented in the output presentation area 253. Thus, the user can determine whether the template of the target prompt word needs to be adjusted based on the model output presented in the output presentation area 253, and by repeating such verification operations for many times, a template of the prompt word that meets the user's needs can be obtained, which can improve the accuracy of the prompt word finally generated.

As shown in FIG2D , an evaluation control 243 may also be presented in the editing control presentation area 240. In some embodiments, in response to detecting a trigger operation on the evaluation control 243, it may be determined that an evaluation request for the target prompt word is received. In response to receiving an evaluation request for the target prompt word, an evaluation input area may be presented. The evaluation input area may be used to receive evaluation data for the target prompt word, where the evaluation data indicates multiple values of at least one variable involved in the template of the target prompt word. Further, the evaluation result of the target prompt word is generated based on the evaluation data. As shown in FIG2D , in response to receiving a trigger operation on the evaluation control 243, it may be determined that an evaluation request for the target prompt word is received. In response to receiving the evaluation request, an evaluation prompt information 260 may be presented in the editing interface 200B. The evaluation prompt information 260 may prompt the user that an evaluation set template for evaluation is being generated by presenting a text such as "generating an evaluation set template". In response to the completion of the generation of the evaluation set template, the presentation of the evaluation prompt information 260 may be cancelled. Furthermore, an evaluation input area 270 as shown in FIG. 2E may be presented.

As shown in Figure 2E, in response to receiving a trigger operation for the evaluation control shown in Figure 2D, it is determined that an evaluation request for the target prompt word is received. In response to receiving the evaluation request, the editing interface 200B can present an evaluation input area 270. The evaluation input area 270 of the model configuration area 230 can, for example, at least include an evaluation set input entrance 271, an evaluation set upload control 272, and an evaluation control 273. The evaluation set input entrance 271 can be used to receive evaluation data (also referred to as an evaluation set) for the target prompt word. The evaluation data here can indicate multiple values of at least one variable involved in the template of the target prompt word. For example, the evaluation data can indicate that the three values for variable A are a, b, and c respectively.

In some embodiments, the evaluation data required for the evaluation can be manually input by the user in the evaluation set input entry 271. In some embodiments, in response to detecting a trigger operation on the evaluation set upload control 272, a selection page for existing evaluation data can be presented, and then in response to the user's selection of a certain evaluation data, the target evaluation data required for the evaluation can be determined. In some embodiments, the evaluation set input entry 271 can also be used to present The evaluation data input by the user or the target evaluation data selected by the user through the evaluation set upload control 272. As shown in FIG2F, the evaluation set input entry 271 may present the evaluation data to be used for evaluation: evaluation set A.

In some embodiments, after the evaluation data required for the evaluation is determined, the evaluation for the target prompt word can be performed in response to detecting a trigger operation on the evaluation control 273. In some embodiments, in response to detecting a trigger operation on the evaluation control 273, the editing interface 200B can present an evaluation prompt area 280 as shown in FIG. 2G, and the evaluation prompt area 180 is used to prompt the user that the evaluation for the target prompt word is currently being performed. In some embodiments, the evaluation prompt area 280 may include a cancel control 281. In response to detecting a trigger operation on the cancel control 281, the presentation of the evaluation prompt area 280 is canceled. In some embodiments, while canceling the presentation of the evaluation prompt area 280, the template of the target prompt word is also stopped from being evaluated.

In some embodiments, in response to detecting a trigger operation on the cancel control 281, only canceling the presentation of the evaluation prompt area 280 will not affect the execution of the evaluation. In this case, after the evaluation prompt area 280 is canceled, the editing interface 200B can present the evaluation prompt information in the evaluation input area 270. As shown in Figure 2H, information such as "evaluation is running" can be presented at the evaluation set input entrance 271 to prompt the user that the evaluation for the target prompt word is currently being performed. In addition, the evaluation termination control 274 can also be presented in the evaluation input area 270. In response to detecting a trigger operation on the evaluation termination control 274, stop evaluating the template of the target prompt word. It can be understood that after the evaluation stops, the information such as "evaluation is running" presented at the evaluation set input entrance 271 will also be stopped from being presented.

Regarding the specific way of executing the evaluation, in some embodiments, the evaluation result can be generated based on the evaluation data for the target prompt word. Specifically, by using the multiple values of at least one variable in the evaluation data to assign at least one variable, multiple test prompt words are generated. For example, assigning the variable A in the template to a, b, and c can obtain three different test prompt words. These multiple test prompt words will be input into the model respectively so that the model generates model outputs for the test prompt words based on these multiple test words. Multiple model outputs for the test prompt words are obtained from the model, and the evaluation result is generated based on the multiple model outputs. In some embodiments, an access entry for the evaluation result can also be provided in the editing interface 200B. As shown in Figure 2I, in response to receiving multiple output results from the model, the evaluation result can be generated based on multiple model outputs. In response to the evaluation result being generated, the editing interface 200B can present an access entry 275 for the evaluation result in the evaluation input area 270. The access entry 275 can be used to present the generated evaluation result. In response to receiving a trigger operation on the evaluation result presented in the access entry 275, the detailed content of the evaluation result can be presented. In addition, a prompt message such as "the evaluation result has been updated in this document" can be presented at the access entry 275 to prompt the user that the evaluation result has been generated and can be viewed. At this point, it can be regarded as completing an evaluation of the target prompt word. Therefore, the prompt word to be released can be evaluated using an evaluation set containing a large number of data sets to determine whether the quality of the prompt word meets expectations, which helps to improve the quality of the prompt word and then improve the quality of subsequent execution tasks.

It should be noted that in the above-mentioned verification and/or evaluation of the target prompt words, the prompt word management system 120 can be integrated with other tools and applications to realize an automated workflow, that is, it can automatically perform verification and/or evaluation of the target prompt words in response to the user's triggering operation on the corresponding control, which can reduce the required manpower and help improve the efficiency of verification and/or evaluation.

It should be noted that, although not shown in FIG. 2E , the editing interface 200B can still include the task definition area 210 and the model configuration area 230, and the same is true for the subsequent figures. In some embodiments, the editing interface 200B also includes a slide control 206. In response to detecting a pull-up operation or a slide-down operation on the slide control 206, Other content is presented in the editing interface 200B. It is understood that other content may be presented in the editing interface 200B in response to detecting a trigger operation on a corresponding physical control and/or in response to detecting a sliding operation in the editing interface 200B (for example, in response to detecting a sliding operation, presenting the task definition area 210).

For the target prompt word of completing verification and/or completing evaluation, in response to receiving prompt word publishing request, the target prompt word is packaged into a callable target task. In certain embodiments, in response to the completion of verification and/or evaluation, editing interface 200B will present publishing control. In response to detecting the triggering operation to publishing control, it is determined that the prompt word publishing request is received. In certain embodiments, the publishing control can only be presented in the editing interface 200B when completing verification and/or completing evaluation. Alternatively or additionally, in certain embodiments, the publishing control can also be presented in the editing interface 200B all the time, even if the target prompt word does not complete verification and/or complete evaluation, it can still be determined in response to detecting the triggering operation to publishing control, and the prompt word publishing request is received, and then the target prompt word is packaged into a callable target task.

As shown in FIG2J , a publishing control 244 may also be presented in the editing control presentation area 240 of the editing interface 200B. It should be noted that, although the evaluation input area 270 is not shown in FIG2J , when the evaluation is completed, the evaluation input area 270 may be presented in the editing interface 200B in response to receiving a pull-down operation on the sliding control 206, in response to detecting a trigger operation on the corresponding physical control, and/or in response to detecting a slide-up operation in the editing interface 200B. In response to detecting a trigger operation on the publishing control 244, it may be determined that a prompt word publishing request has been received.

In some embodiments, the user can also select the version number of the target task to be packaged before triggering the release control. As shown in Figure 2I, the editing interface 200B can also present a version selection entry 207. In response to detecting the trigger operation of the version selection entry 207, at least one version number can be presented, for example, the four version numbers "$DEV, 3, 2, 1" are presented. It should be noted that when the version number selection entry 207 corresponds to the development mark DEV, it means that it is currently in the development state, that is, it is in the state of editing and unpublishing the target prompt word. In this case, the target prompt word cannot be called for a task, and only operations such as editing, verification, and evaluation can be performed on the target prompt word. When the version number selection entry 207 corresponds to a version number other than DEV, for example, it corresponds to 1, which means that the target task is associated with version number 1.

In some embodiments, when it is determined that the target task is associated with a specific version number that is not a development identifier, the editing interface 200B may also present a version disable identifier 208. In response to detecting a trigger operation on the version disable identifier 208, the currently selected version number is disabled. In some embodiments, the editing interface 200B also presents a status identifier control 209, which is used to indicate the current version status. Exemplarily, the status identifier control 209 may present three states: DEV, Activated, and Disabled.

It should be noted that, regarding the version number associated with the target task, when the prompt word edit request is a task creation request, in response to the prompt word release request, the target task is associated with the first version number. The first version number here can be any version number. When the prompt word edit request is a modification request for a created task, in response to the prompt word release request, the target task is associated with the second version number. The second version number here is different from the version number of the created task. Exemplarily, in the creation stage of the target task, the version number associated with the target task can be 1, and in the modification stage of the target task, the version number associated with the modified target task can be 2, 3, 4, etc. Thus, target tasks at different stages can be associated with different version numbers, which helps users to search for target tasks later. For example, if a user expects to find the target task after the second modification, the user can directly find the target task after the second modification from multiple target tasks based on the version number associated with the target task after the second modification.

In some embodiments, since target tasks at different stages are associated with different version numbers (e.g., target tasks at the creation stage and target tasks at the first modification stage), the target task can also be called based on the version number of the target task. Thus, the target task to be called can be determined from multiple target tasks conveniently and quickly, and target tasks of different historical versions can be called, which helps to improve the efficiency of target task calling.

Referring to Fig. 3, Fig. 3 shows a schematic diagram of a process 300 of prompt word management according to some embodiments of the present disclosure. The process 300 may be implemented at the prompt word management system 120. For ease of discussion, the process 300 will be described with reference to the environment 100 of Fig. 1 .

At block 310 , the cue word management system 120 creates a target task on the cue word management system 120 in response to the received user input.

In block 320 , the prompt word management system 120 defines a prompt word template in response to the user input in the prompt word management system 120 . The prompt word template here is a template of the target prompt word.

In block 330 , in response to receiving the evaluation request for the prompt word, the prompt word management system 120 performs an evaluation operation on the prompt word to evaluate the actual effect of the prompt word.

In block 340, it is determined whether the prompt word has reached the online standard. In some embodiments, it is possible to determine whether the prompt word meets the user's expectations based on the actual effect of the evaluation prompt word, and if it is determined to meet the expectations, it is determined that the online standard has been reached. Further, in block 350, the prompt word that has reached the online standard can be released.

In some embodiments, if the actual effect obtained by the evaluation does not meet expectations, it can be determined that the online standard is not met. The prompt word management system 120 can prompt the user to modify the prompt word template. That is, return to the step shown in the execution box 320. In some embodiments, the evaluation shown in the box 330 and the judgment step shown in the box 340 can be continued for the modified prompt word template until the modified prompt word template meets the online standard and can be released.

In some embodiments, in addition to executing the evaluation method shown in block 33 on the prompt word, verification may also be performed on it to determine its actual effect.

In summary, according to the embodiments of the present disclosure, prompt words can be flexibly designed and managed, and prompt words can be encapsulated into callable tasks, which makes the calling of the model very simple and can improve the stability, reliability, convenience and efficiency of prompt word management and model calling in practical applications.

It should be understood that some embodiments of the present disclosure are described above in conjunction with the specific examples in the drawings, but these specific examples do not limit the scope of the embodiments of the present disclosure. The described embodiments can also be implemented in various other variations.

FIG4 shows a flow chart of a process 400 for prompt word management according to some embodiments of the present disclosure. The process 400 may be implemented at the prompt word management system 120. The process 400 is described below with reference to FIG1.

In block 410 , in response to receiving a prompt word editing request, the prompt word management system 120 presents an editing interface, wherein the editing interface at least includes an input area for receiving a target prompt word.

In block 420 , in response to receiving the prompt word publishing request, the prompt word management system 120 encapsulates the target prompt word into a callable target task based at least on the user input in the editing interface, wherein calling the target task will trigger the target prompt word to be input into the model.

In some embodiments, the editing interface also includes a variable definition area for receiving a definition of at least one variable involved in the target prompt word, and wherein a call to the target task will trigger at least one variable to be assigned a value and the target prompt word with at least one assigned variable to be input into the model.

In some embodiments, the editing interface further comprises a model definition area for receiving a selection of a target model, and wherein the invocation of the target task triggers the target prompt word to be input into the selected target model.

In some embodiments, the model definition area is further used to receive a definition of processing parameters of the target model, and wherein the call to the target task will trigger the target model to process the target prompt word according to the defined processing parameters.

In some embodiments, the editing interface further comprises a task definition area for receiving a task identifier of the target task, and wherein the calling of the target task comprises calling the target task according to the task identifier of the target task.

In some embodiments, the input area is used to receive a template of a target prompt word, and the template defines the structure of the target prompt word.

In some embodiments, process 400 also includes: before receiving a prompt word release request, in response to receiving a verification request for a target prompt word, inputting the target prompt word into the model based at least on user input in the editing interface; receiving a model output for the target prompt word from the model; and presenting the received model output.

In some embodiments, process 400 also includes: in the case where the prompt word edit request is a task creation request, in response to the prompt word publish request, associating the target task to a first version number; and in the case where the prompt word edit request is a modification request for an already created task, in response to the prompt word publish request, associating the target task to a second version number, the second version number being different from the version number of the already created task.

In some embodiments, calling the target task includes calling the target task according to a version number of the target task.

In some embodiments, process 400 also includes: in response to receiving an evaluation request for a target prompt word, presenting an evaluation input area; receiving evaluation data for the target prompt word in the evaluation input area, the evaluation data indicating multiple values of at least one variable involved in the target prompt word; generating an evaluation result for the target prompt word based on the evaluation data; and providing access to the evaluation result.

In some embodiments, generating an evaluation result includes: generating multiple test prompt words by assigning at least one variable using multiple values for at least one variable in the evaluation data respectively; inputting the multiple test prompt words into the model respectively; receiving multiple model outputs for the test prompt words from the model; and generating an evaluation result based on the multiple model outputs.

In some embodiments, presenting the editing interface includes: determining the authority of the user who initiated the prompt word editing request if the prompt word editing request is a modification request for an already created task; and presenting the editing interface if the user has an editing request for the already created task.

In some embodiments, the process 400 further includes: presenting a task list on the prompt word management system, the task list including at least one created task.

5 shows a schematic structural block diagram of an apparatus 500 for prompt word management according to some embodiments of the present disclosure. The apparatus 500 may be implemented as or included in the prompt word management system 120. Each module/component in the apparatus 500 may be implemented by hardware, software, firmware or any combination thereof.

As shown in the figure, the device 500 includes an interface presentation module 510, which is configured to present an editing interface in response to receiving a prompt word editing request, wherein the editing interface at least includes an input area for receiving a target prompt word. The device 700 also includes a publishing module 520, which is configured to encapsulate the target prompt word into a callable target task in response to receiving a prompt word publishing request, at least based on user input in the editing interface, wherein the call of the target task will trigger the target prompt word to be input into the model.

In some embodiments, the editing interface also includes a variable definition area for receiving a definition of at least one variable involved in the target prompt word, and wherein a call to the target task will trigger at least one variable to be assigned a value and the target prompt word with at least one assigned variable to be input into the model.

In some embodiments, the editing interface further comprises a model definition area for receiving a selection of a target model, and wherein the invocation of the target task triggers the target prompt word to be input into the selected target model.

In some embodiments, the model definition area is further used to receive a definition of processing parameters of the target model, and wherein the call to the target task will trigger the target model to process the target prompt word according to the defined processing parameters.

In some embodiments, the editing interface further comprises a task definition area for receiving a task identifier of the target task, and wherein the calling of the target task comprises calling the target task according to the task identifier of the target task.

In some embodiments, the input area is used to receive a template of a target prompt word, and the template defines the structure of the target prompt word.

In some embodiments, the device 500 also includes: a prompt word input module, which is configured to input the target prompt word into the model in response to receiving a verification request for the target prompt word before receiving the prompt word release request, at least based on the user input in the editing interface; an output receiving module, which is configured to receive the model output for the target prompt word from the model; and an output presentation module, which is configured to present the received model output.

In some embodiments, the device 500 also includes: a first association module, configured to associate the target task with a first version number in response to a prompt word publication request when the prompt word editing request is a task creation request; and a second association module, configured to associate the target task with a second version number in response to the prompt word publication request when the prompt word editing request is a modification request for an already created task, the second version number being different from the version number of the already created task.

In some embodiments, calling the target task includes calling the target task according to a version number of the target task.

In some embodiments, the device 500 also includes: an evaluation area presentation module, configured to present an evaluation input area in response to receiving an evaluation request for a target prompt word; an evaluation data receiving module, configured to receive evaluation data for the target prompt word in the evaluation input area, the evaluation data indicating multiple values of at least one variable involved in the target prompt word; an evaluation result generation module, configured to generate an evaluation result for the target prompt word based on the evaluation data; and an access entry providing module, configured to provide an access entry for the evaluation result.

In some embodiments, the evaluation result generation module includes: a test prompt word generation module, configured to generate multiple test prompt words by assigning at least one variable by respectively using multiple values for at least one variable in the evaluation data; an input module, configured to input the multiple test prompt words into the model respectively; a receiving module, configured to receive multiple model outputs for the test prompt words from the model; and a generation module, configured to generate evaluation results based on the multiple model outputs.

In some embodiments, the interface presentation module 510 includes: an authority determination module, configured to determine the authority of the user who initiates the prompt word editing request when the prompt word editing request is a modification request for an already created task; and a presentation module, configured to present an editing interface if the user has an editing request for an already created task.

In some embodiments, the apparatus 500 further includes: a task list presenting module configured to present a task list on the prompt word management system, wherein the task list includes at least one created task.

The modules and/or units included in the apparatus 500 may be implemented in various ways, including software, hardware, firmware, or any combination thereof. In some embodiments, one or more modules and/or units may be implemented using software and/or firmware, such as machine executable instructions stored on a storage medium. Alternatively, some or all of the modules and/or units in the device 500 may be at least partially implemented by one or more hardware logic components. By way of example and not limitation, exemplary types of hardware logic components that may be used include field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

It should be understood that one or more steps in the above method may be performed by an appropriate electronic device or combination of electronic devices, which may include, for example, the electronic device 110 in FIG. 1 and/or a combination with other electronic devices.

FIG6 shows a block diagram of an electronic device 600 in which one or more embodiments of the present disclosure may be implemented. It should be understood that the electronic device 600 shown in FIG6 is merely exemplary and should not constitute any limitation on the functionality and scope of the embodiments described herein. The electronic device 600 shown in FIG6 may be used to implement the electronic device 110 of FIG1 .

As shown in FIG6 , the electronic device 600 is in the form of a general electronic device. The components of the electronic device 600 may include, but are not limited to, one or more processors or processing units 610, a memory 620, a storage device 630, one or more communication units 640, one or more input devices 650, and one or more output devices 660. The processing unit 610 may be an actual or virtual processor and is capable of performing various processes according to a program stored in the memory 620. In a multi-processor system, multiple processing units execute computer executable instructions in parallel to improve the parallel processing capability of the electronic device 600.

The electronic device 600 typically includes a plurality of computer storage media. Such media can be any available media accessible to the electronic device 600, including but not limited to volatile and non-volatile media, removable and non-removable media. The memory 620 can be a volatile memory (e.g., registers, caches, random access memory (RAM)), a non-volatile memory (e.g., a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), flash memory) or some combination thereof. The storage device 630 can be a removable or non-removable medium, and can include a machine-readable medium, such as a flash drive, a disk, or any other medium, which can be used to store information and/or data and can be accessed within the electronic device 600.

The electronic device 600 may further include additional removable/non-removable, volatile/non-volatile storage media. Although not shown in FIG. 6 , a disk drive for reading or writing from a removable, non-volatile disk (e.g., a “floppy disk”) and an optical drive for reading or writing from a removable, non-volatile optical disk may be provided. In these cases, each drive may be connected to a bus (not shown) by one or more data media interfaces. The memory 620 may include a computer program product 625 having one or more program modules configured to perform various methods or actions of various embodiments of the present disclosure.

The communication unit 640 implements communication with other electronic devices through a communication medium. Additionally, the functions of the components of the electronic device 600 can be implemented with a single computing cluster or multiple computing machines that can communicate through a communication connection. Therefore, the electronic device 600 can operate in a networked environment using a logical connection with one or more other servers, a network personal computer (PC), or another network node.

The input device 650 may be one or more input devices, such as a mouse, a keyboard, a tracking ball, etc. The output device 660 may be one or more output devices, such as a display, a speaker, a printer, etc. The electronic device 600 may also communicate with one or more external devices (not shown) through the communication unit 640 as needed, such as a storage device, a display device, etc., communicate with one or more devices that allow a user to interact with the electronic device 600, or communicate with any device that allows the electronic device 600 to communicate with one or more other electronic devices (e.g., Such communication can be performed via an input/output (I/O) interface (not shown).

According to an exemplary implementation of the present disclosure, a computer-readable storage medium is provided, on which computer-executable instructions are stored, wherein the computer-executable instructions are executed by a processor to implement the method described above. According to an exemplary implementation of the present disclosure, a computer program product is also provided, which is tangibly stored on a non-transitory computer-readable medium and includes computer-executable instructions, and the computer-executable instructions are executed by a processor to implement the method described above.

Various aspects of the present disclosure are described herein with reference to the flowcharts and/or block diagrams of the methods, devices, equipment, and computer program products implemented according to the present disclosure. It should be understood that each box in the flowchart and/or block diagram and the combination of each box in the flowchart and/or block diagram can be implemented by computer-readable program instructions.

These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processing unit of the computer or other programmable data processing device, a device that implements the functions/actions specified in one or more boxes in the flowchart and/or block diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and/or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions/actions specified in one or more boxes in the flowchart and/or block diagram.

Computer-readable program instructions can be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, so that the instructions executed on the computer, other programmable data processing apparatus, or other device implement the functions/actions specified in one or more boxes in the flowchart and/or block diagram.

The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to multiple implementations of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of a module, program segment or instruction includes one or more executable instructions for realizing the logical function of the specification. In some implementations as replacements, the function marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each square box in the block diagram and/or flow chart, and the combination of the square boxes in the block diagram and/or flow chart can be realized by a special hardware-based system that performs the function or action of the specification, or can be realized by a combination of special hardware and computer instructions.

The above descriptions of various implementations of the present disclosure are exemplary, non-exhaustive, and not limited to the disclosed implementations. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described implementations. The selection of terms used herein is intended to best explain the principles of the implementations, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the various implementations disclosed herein.

Claims (16)

A method for prompt word management, executed by a prompt word management system, the method comprising: In response to receiving a prompt word editing request, presenting an editing interface, wherein the editing interface at least includes an input area for receiving a target prompt word; and In response to receiving a prompt word publishing request, the target prompt word is encapsulated into a callable target task based at least on user input in the editing interface, wherein calling the target task will trigger the target prompt word to be input into the model. The method according to claim 1, wherein the editing interface further comprises a variable definition area for receiving a definition of at least one variable involved in the target prompt word, and The call to the target task will trigger the at least one variable to be assigned a value and the target prompt word with the at least one assigned variable to be input into the model. The method according to claim 1, wherein the editing interface further comprises a model definition area for receiving a selection of a target model, and The call to the target task will trigger the target prompt word to be input into the selected target model. The method according to claim 3, wherein the model definition area is further used to receive a definition of processing parameters of the target model, and The call to the target task will trigger the target model to process the target prompt word according to the defined processing parameters. The method according to claim 1, wherein the editing interface further comprises a task definition area for receiving a task identifier of the target task, and The calling of the target task includes calling the target task according to a task identifier of the target task. The method according to claim 1, wherein the input area is used to receive a template of the target prompt word, and the template defines the structure of the target prompt word. The method according to claim 1, further comprising: Before receiving the prompt word publishing request, in response to receiving a verification request for the target prompt word, inputting the target prompt word into the model based at least on a user input in the editing interface; receiving a model output for the target cue word from the model; and The received model output is presented. The method according to claim 1, further comprising: In the case where the prompt word editing request is a task creation request, in response to the prompt word publishing request, associating the target task with a first version number; and In the case that the prompt word editing request is a modification request for a created task, in response to the prompt word publishing request, the target task is associated with a second version number that is different from the version number of the created task. The method according to claim 8, wherein calling the target task comprises calling the target task according to a version number of the target task. The method according to claim 1, further comprising: In response to receiving an evaluation request for the target prompt word, presenting an evaluation input area; receiving evaluation data for the target prompt word in the evaluation input area, the evaluation data indicating multiple values of at least one variable involved in the target prompt word; generating an evaluation result for the target prompt word based on the evaluation data; and Providing access to the evaluation results. The method according to claim 10, wherein generating the evaluation result comprises: Generating a plurality of test prompt words by respectively assigning a value to the at least one variable using a plurality of values for the at least one variable in the evaluation data; Inputting the plurality of test prompt words into the model respectively; receiving a plurality of model outputs from the model for the test cue word; and The evaluation result is generated based on the plurality of model outputs. The method according to claim 1, wherein presenting the editing interface comprises: In the case where the prompt word editing request is a request to modify a created task, determining the authority of the user who initiated the prompt word editing request; and If the user has an editing request for the created task, the editing interface is presented. The method according to claim 1, further comprising: A task list is presented on the prompt word management system, wherein the task list includes at least one created task. A device for prompt word management, comprising: an interface presenting module, configured to present an editing interface in response to receiving a prompt word editing request, wherein the editing interface at least includes an input area for receiving a target prompt word; and The publishing module is configured to encapsulate the target prompt word into a callable target task in response to receiving a prompt word publishing request, at least based on user input in the editing interface, wherein the call to the target task will trigger the target prompt word to be input into the model. An electronic device, comprising: at least one processing unit; and At least one memory, the at least one memory being coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit, the instructions causing the electronic device to perform the method according to any one of claims 1 to 13 when executed by the at least one processing unit. A computer-readable storage medium having a computer program stored thereon, wherein the computer program can be executed by a processor to implement the method according to any one of claims 1 to 13.