{"id":18129,"date":"2026-09-22T05:36:21","date_gmt":"2026-09-22T05:36:21","guid":{"rendered":"https:\/\/www.examlabs.com\/certification\/?p=18129"},"modified":"2026-09-22T05:36:21","modified_gmt":"2026-09-22T05:36:21","slug":"databricks-certified-generative-ai-engineer-associate-practice-test-questions-and-exam-dumps-part4-q61-80","status":"publish","type":"post","link":"https:\/\/www.examlabs.com\/certification\/databricks-certified-generative-ai-engineer-associate-practice-test-questions-and-exam-dumps-part4-q61-80\/","title":{"rendered":"Databricks Certified Generative AI Engineer Associate Practice Test Questions and Exam Dumps Part4 Q61-80"},"content":{"rendered":"

View Full <\/b>Databricks Certified Generative AI Engineer Associate Exam Dumps<\/b><\/a> and Practice Test Dumps<\/b><\/p>\n

 <\/p>\n

Question 61. What is the primary purpose of prompt templates in a generative AI application?<\/b><\/p>\n

    \n
  1. <\/b> To provide a reusable structure for constructing model inputs<\/span><\/li>\n
  2. <\/b> To replace vector databases<\/span><\/li>\n
  3. <\/b> To increase physical storage capacity<\/span><\/li>\n
  4. <\/b> To manage network routing<\/span><\/li>\n<\/ol>\n

    Correct Answer: 1. To provide a reusable structure for constructing model inputs<\/b><\/p>\n

    Explanation:<\/b><\/p>\n

    Prompt templates provide a consistent structure for creating prompts sent to a generative AI model. They can contain fixed instructions, placeholders for user input, retrieved context, conversation history, and other dynamic information. Using templates helps developers maintain consistency across requests and makes it easier to update application instructions without rewriting the entire application logic. In a RAG application, a template may include instructions telling the model how to use retrieved documents when generating an answer. Templates do not replace vector databases or manage network routing. They are an application-development technique that supports reusable, maintainable, and more predictable prompt construction across different requests and use cases.<\/span><\/p>\n

    Question 62. What is the purpose of retrieving documents before generating a response in a RAG system?<\/b><\/p>\n

      \n
    1. <\/b> To permanently modify the foundation model<\/span><\/li>\n
    2. <\/b> To provide relevant information that can ground the generated response<\/span><\/li>\n
    3. <\/b> To remove the need for evaluation<\/span><\/li>\n
    4. <\/b> To increase database storage automatically<\/span><\/li>\n<\/ol>\n

      Correct Answer: 2. To provide relevant information that can ground the generated response<\/b><\/p>\n

      Explanation:<\/b><\/p>\n

      Retrieving documents before generation allows a RAG application to provide the language model with external information relevant to the user’s question. The retrieved content becomes part of the model’s context and can help the model produce an answer based on current or organization-specific information. This is especially useful when the knowledge required by the application is not reliably contained in the model’s parameters. Retrieval does not permanently modify the foundation model and does not eliminate the need for evaluation. The quality of the final response depends partly on retrieval quality, so applications should evaluate whether the retrieved documents are relevant, sufficiently complete, and appropriate for the query before relying on them as grounding context.<\/span><\/p>\n

      Question 63. What is an important consideration when selecting an embedding model for a RAG application?<\/b><\/p>\n

        \n
      1. <\/b> The physical location of the user’s computer<\/span><\/li>\n
      2. <\/b> Whether it supports the required language and semantic retrieval task<\/span><\/li>\n
      3. <\/b> The color of the application’s user interface<\/span><\/li>\n
      4. <\/b> The number of SQL dashboards<\/span><\/li>\n<\/ol>\n

        Correct Answer: 2. Whether it supports the required language and semantic retrieval task<\/b><\/p>\n

        Explanation:<\/b><\/p>\n

        Embedding model selection can have a direct impact on retrieval quality because embeddings determine how text is represented for semantic similarity comparison. Teams should consider whether the embedding model performs well for the languages, terminology, and types of content used by the application. Other factors can include embedding dimensions, inference cost, latency, supported input types, and compatibility with the selected vector search system. A model that performs well on general text may not necessarily provide optimal retrieval for specialized domains. Evaluation using representative queries and documents is therefore important. Interface appearance and SQL dashboard count do not determine whether an embedding model is suitable for a RAG workload.<\/span><\/p>\n

        Question 64. What is semantic similarity used for in vector retrieval?<\/b><\/p>\n

          \n
        1. <\/b> Comparing representations to identify conceptually related content<\/span><\/li>\n
        2. <\/b> Encrypting database passwords<\/span><\/li>\n
        3. <\/b> Creating network policies<\/span><\/li>\n
        4. <\/b> Measuring physical storage capacity<\/span><\/li>\n<\/ol>\n

          Correct Answer: 1. Comparing representations to identify conceptually related content<\/b><\/p>\n

          Explanation:<\/b><\/p>\n

          Semantic similarity measures how closely two representations are related in meaning. In vector retrieval, a query is typically converted into an embedding and compared with embeddings representing stored documents or chunks. Content with vectors that are sufficiently similar can be considered potentially relevant to the query. This differs from exact keyword matching because semantically related text may use different words while still expressing similar concepts. Semantic similarity is therefore an important foundation of vector-based RAG retrieval. The exact similarity calculation and ranking method depend on the retrieval system. Encryption, network policies, and storage capacity are separate infrastructure concerns and do not perform semantic comparison between document representations.<\/span><\/p>\n

          Question 65. Why is source data quality important for a RAG application?<\/b><\/p>\n

            \n
          1. <\/b> Poor source data can lead to inaccurate or misleading retrieved context<\/span><\/li>\n
          2. <\/b> Source quality only affects physical storage costs<\/span><\/li>\n
          3. <\/b> Source data has no relationship to generated answers<\/span><\/li>\n
          4. <\/b> High-quality data eliminates all model evaluation<\/span><\/li>\n<\/ol>\n

            Correct Answer: 1. Poor source data can lead to inaccurate or misleading retrieved context<\/b><\/p>\n

            Explanation:<\/b><\/p>\n

            The quality of information available to a RAG system directly influences the quality of the context retrieved for model generation. If source documents contain outdated, incorrect, duplicated, or poorly structured information, the retrieval system may return content that leads the model toward an inaccurate answer. Data preparation can therefore include cleaning documents, removing unnecessary duplicates, maintaining current versions, validating content, and preserving useful metadata. High-quality source data does not guarantee perfect responses because retrieval and generation can still introduce errors. However, improving the underlying knowledge source establishes a stronger foundation for reliable retrieval and helps the model receive information that is more appropriate for the application’s intended domain.<\/span><\/p>\n

            Question 66. What is one reason to remove duplicate or obsolete documents from a RAG knowledge base?<\/b><\/p>\n

              \n
            1. <\/b> To make retrieval results more consistent and reduce conflicting information<\/span><\/li>\n
            2. <\/b> To increase model temperature<\/span><\/li>\n
            3. <\/b> To disable semantic search<\/span><\/li>\n
            4. <\/b> To permanently change model weights<\/span><\/li>\n<\/ol>\n

              Correct Answer: 1. To make retrieval results more consistent and reduce conflicting information<\/b><\/p>\n

              Explanation:<\/b><\/p>\n

              Duplicate and obsolete documents can create problems in a retrieval system because multiple versions of similar information may be returned for the same query. If outdated information conflicts with a newer policy or document, the language model may receive ambiguous context and generate an incorrect response. Maintaining a clean knowledge base can therefore improve retrieval quality and reduce unnecessary noise. Document lifecycle management should identify authoritative versions and remove or appropriately label outdated material. This does not involve changing model weights or disabling semantic search. Data quality is an important part of RAG engineering because the generative model depends heavily on the information supplied by the retrieval layer.<\/span><\/p>\n

              Question 67. What does top-k retrieval typically specify?<\/b><\/p>\n

                \n
              1. <\/b> The number of highest-ranked results returned by a retrieval operation<\/span><\/li>\n
              2. <\/b> The number of model parameters<\/span><\/li>\n
              3. <\/b> The number of GPUs in a cluster<\/span><\/li>\n
              4. <\/b> The number of users in an application<\/span><\/li>\n<\/ol>\n

                Correct Answer: 1. The number of highest-ranked results returned by a retrieval operation<\/b><\/p>\n

                Explanation:<\/b><\/p>\n

                Top-k retrieval refers to returning the k highest-ranked results from a retrieval operation. For example, a system configured with a top-k value of five may return the five highest-ranked document chunks according to its retrieval scoring mechanism. The selected value affects the amount of context provided to the generative model. A very small value may omit useful information, while a very large value may introduce irrelevant content and consume more context-window capacity. Therefore, top-k should be tuned and evaluated according to the application’s retrieval requirements. It does not represent model parameter count, GPU count, or user count. Retrieval evaluation can help determine an appropriate value.<\/span><\/p>\n

                Question 68. What is the purpose of a similarity threshold in vector retrieval?<\/b><\/p>\n

                  \n
                1. <\/b> To determine whether retrieved results are sufficiently similar to the query<\/span><\/li>\n
                2. <\/b> To control physical server temperature<\/span><\/li>\n
                3. <\/b> To increase the number of model layers<\/span><\/li>\n
                4. <\/b> To create user accounts<\/span><\/li>\n<\/ol>\n

                  Correct Answer: 1. To determine whether retrieved results are sufficiently similar to the query<\/b><\/p>\n

                  Explanation:<\/b><\/p>\n

                  A similarity threshold can be used to filter retrieval results based on their similarity score to the query. If a candidate result falls below the configured threshold, the application may choose not to include it in the retrieved context. This can help reduce irrelevant information being passed to the language model. However, setting the threshold too high may exclude useful information, while setting it too low may allow weakly relevant content into the context. The appropriate threshold should therefore be evaluated using representative queries and retrieval results. Similarity thresholds are retrieval controls and are unrelated to server temperature, model layer count, or user-account creation.<\/span><\/p>\n

                  Question 69. What is the main goal of a RAG evaluation pipeline?<\/b><\/p>\n

                    \n
                  1. <\/b> To measure retrieval and generation quality against defined criteria<\/span><\/li>\n
                  2. <\/b> To automatically increase model size<\/span><\/li>\n
                  3. <\/b> To remove all source documents<\/span><\/li>\n
                  4. <\/b> To eliminate production monitoring<\/span><\/li>\n<\/ol>\n

                    Correct Answer: 1. To measure retrieval and generation quality against defined criteria<\/b><\/p>\n

                    Explanation:<\/b><\/p>\n

                    A RAG evaluation pipeline provides a systematic way to assess how well the retrieval and generation components perform. It can evaluate whether relevant documents are retrieved, whether generated answers are grounded in those documents, and whether responses satisfy application-specific quality requirements. Evaluation may include automated metrics, human review, or a combination of both. Running the same evaluation dataset against different configurations allows teams to compare changes to embeddings, chunking, retrieval parameters, prompts, or models. Evaluation does not automatically increase model size or eliminate production monitoring. Instead, it provides evidence that can guide engineering decisions and help identify regressions before and after application changes.<\/span><\/p>\n

                    Question 70. Which metric is commonly associated with the relevance of a generated answer to the user’s question?<\/b><\/p>\n

                      \n
                    1. <\/b> Response relevance<\/span><\/li>\n
                    2. <\/b> CPU temperature<\/span><\/li>\n
                    3. <\/b> Disk capacity<\/span><\/li>\n
                    4. <\/b> Network packet size<\/span><\/li>\n<\/ol>\n

                      Correct Answer: 1. Response relevance<\/b><\/p>\n

                      Explanation:<\/b><\/p>\n

                      Response relevance measures how well a generated answer addresses the user’s actual question or task. A response may be grammatically correct and well written but still be irrelevant if it does not address what the user asked. Relevance evaluation can be performed using automated evaluators, model-based judges, human reviewers, or combinations of these approaches depending on the application. In a RAG system, relevance should be considered alongside other dimensions such as groundedness, correctness, safety, and completeness. Infrastructure measurements such as disk capacity and CPU temperature do not directly assess the usefulness of a generated answer. Evaluating relevance helps determine whether the application is meeting its intended conversational or informational purpose.<\/span><\/p>\n

                      Question 71. What does groundedness evaluation attempt to determine?<\/b><\/p>\n

                        \n
                      1. <\/b> Whether generated claims are supported by the provided context or sources<\/span><\/li>\n
                      2. <\/b> Whether the cluster has enough disk space<\/span><\/li>\n
                      3. <\/b> Whether the application has enough user accounts<\/span><\/li>\n
                      4. <\/b> Whether the model has the maximum possible parameter count<\/span><\/li>\n<\/ol>\n

                        Correct Answer: 1. Whether generated claims are supported by the provided context or sources<\/b><\/p>\n

                        Explanation:<\/b><\/p>\n

                        Groundedness evaluation examines whether the claims made in a generated response are supported by the information available to the model. In a RAG application, this generally means checking whether statements in the answer can be traced to the retrieved documents or other approved evidence. A response can be relevant to a question while still containing unsupported claims, so groundedness is a distinct evaluation dimension. This type of evaluation can help identify hallucination risks and weaknesses in retrieval or prompting. Groundedness does not measure cluster capacity or model size. It focuses on the relationship between generated content and the evidence supplied to the application during inference.<\/span><\/p>\n

                        Question 72. Why can human evaluation still be useful for generative AI applications?<\/b><\/p>\n

                          \n
                        1. <\/b> Human reviewers can assess qualitative characteristics that automated metrics may miss<\/span><\/li>\n
                        2. <\/b> Humans automatically increase model context windows<\/span><\/li>\n
                        3. <\/b> Human review replaces all technical testing<\/span><\/li>\n
                        4. <\/b> Human reviewers eliminate the need for monitoring<\/span><\/li>\n<\/ol>\n

                          Correct Answer: 1. Human reviewers can assess qualitative characteristics that automated metrics may miss<\/b><\/p>\n

                          Explanation:<\/b><\/p>\n

                          Human evaluation can provide valuable judgments about qualities that may be difficult to capture completely with automated metrics. Reviewers can assess whether an answer is useful, clear, appropriate, complete, and aligned with the intended task. They may also identify subtle safety, tone, or factual issues that an automated metric fails to detect. Human evaluation does not replace technical testing, monitoring, or automated evaluation. Instead, it can complement those approaches by providing qualitative evidence. For important generative AI applications, combining automated measurements with structured human review can provide a broader understanding of system behavior and help teams identify areas requiring further improvement.<\/span><\/p>\n

                          Question 73. What is the purpose of an evaluation baseline in a generative AI project?<\/b><\/p>\n

                            \n
                          1. <\/b> To provide a reference point for comparing subsequent changes<\/span><\/li>\n
                          2. <\/b> To guarantee that all future versions perform identically<\/span><\/li>\n
                          3. <\/b> To eliminate application logs<\/span><\/li>\n
                          4. <\/b> To prevent any model updates<\/span><\/li>\n<\/ol>\n

                            Correct Answer: 1. To provide a reference point for comparing subsequent changes<\/b><\/p>\n

                            Explanation:<\/b><\/p>\n

                            An evaluation baseline establishes a reference measurement against which future versions of a generative AI application can be compared. A baseline might include retrieval metrics, answer quality, groundedness, safety measurements, latency, or other application-specific criteria. When developers modify prompts, models, retrieval settings, chunking, or other components, they can compare new evaluation results with the baseline to determine whether the change improved or degraded performance. A baseline does not guarantee identical behavior across versions and does not prevent model updates. Its value is that it provides a consistent point of comparison, making it easier to identify meaningful regressions and improvements during iterative development.<\/span><\/p>\n

                            Question 74. Why is evaluation data leakage a concern when testing a generative AI system?<\/b><\/p>\n

                              \n
                            1. <\/b> It can make evaluation results less representative of unseen performance<\/span><\/li>\n
                            2. <\/b> It always increases inference latency<\/span><\/li>\n
                            3. <\/b> It prevents models from generating text<\/span><\/li>\n
                            4. <\/b> It automatically improves retrieval precision<\/span><\/li>\n<\/ol>\n

                              Correct Answer: 1. It can make evaluation results less representative of unseen performance<\/b><\/p>\n

                              Explanation:<\/b><\/p>\n

                              Evaluation data leakage occurs when information from an evaluation set improperly influences the system being evaluated. For example, if test examples or their answers become part of training or optimization data, the resulting evaluation may make the system appear stronger than it would be on genuinely unseen inputs. This reduces confidence that evaluation results represent real-world generalization. Keeping evaluation datasets appropriately separated helps provide a more objective measurement of application behavior. Leakage does not necessarily increase inference latency or prevent text generation. Careful dataset management is therefore important when evaluating prompts, retrieval strategies, fine-tuned models, and complete generative AI applications.<\/span><\/p>\n

                              Question 75. What is one advantage of using an automated evaluation process for repeated model experiments?<\/b><\/p>\n

                                \n
                              1. <\/b> It enables consistent comparison across multiple configurations<\/span><\/li>\n
                              2. <\/b> It guarantees that all generated responses are correct<\/span><\/li>\n
                              3. <\/b> It eliminates the need for test datasets<\/span><\/li>\n
                              4. <\/b> It prevents model changes<\/span><\/li>\n<\/ol>\n

                                Correct Answer: 1. It enables consistent comparison across multiple configurations<\/b><\/p>\n

                                Explanation:<\/b><\/p>\n

                                Automated evaluation allows teams to run the same tests against multiple model, prompt, retrieval, or application configurations. Because the evaluation procedure can be standardized, results can be compared more consistently than when every experiment is reviewed using a different process. This is particularly useful when development involves many iterations and teams need evidence about whether a change improves application quality. Automated evaluation does not guarantee that every generated answer is correct, and it does not eliminate the need for carefully designed test datasets or human review where appropriate. Its main benefit is repeatability and efficiency, allowing teams to measure changes systematically and identify regressions more quickly.<\/span><\/p>\n

                                Question 76. What is a common reason to use synthetic data when developing a generative AI application?<\/b><\/p>\n

                                  \n
                                1. <\/b> To create additional representative test or development examples when real data is limited<\/span><\/li>\n
                                2. <\/b> To guarantee perfect model accuracy<\/span><\/li>\n
                                3. <\/b> To replace all production data permanently<\/span><\/li>\n
                                4. <\/b> To eliminate privacy considerations<\/span><\/li>\n<\/ol>\n

                                  Correct Answer: 1. To create additional representative test or development examples when real data is limited<\/b><\/p>\n

                                  Explanation:<\/b><\/p>\n

                                  Synthetic data can be generated to provide additional examples for development, testing, or evaluation when suitable real-world data is limited or difficult to use. It can help teams create scenarios that represent different questions, edge cases, or application behaviors. However, synthetic data should be validated because generated examples may contain unrealistic patterns, biases, or inaccuracies. It does not guarantee model accuracy, replace production data in every situation, or eliminate privacy considerations. Teams should determine whether synthetic examples adequately represent the intended workload and should avoid assuming that performance on synthetic data automatically predicts real-world performance. Careful dataset design remains important regardless of how examples are produced.<\/span><\/p>\n

                                  Question 77. What is the purpose of a model’s system instruction when building a domain-specific assistant?<\/b><\/p>\n

                                    \n
                                  1. <\/b> To establish the assistant’s role, rules, and response behavior<\/span><\/li>\n
                                  2. <\/b> To store vector embeddings<\/span><\/li>\n
                                  3. <\/b> To increase database capacity<\/span><\/li>\n
                                  4. <\/b> To provision physical servers<\/span><\/li>\n<\/ol>\n

                                    Correct Answer: 1. To establish the assistant’s role, rules, and response behavior<\/b><\/p>\n

                                    Explanation:<\/b><\/p>\n

                                    System instructions provide high-level guidance that can define how an AI assistant should behave. They may establish the assistant’s role, specify response requirements, define boundaries, and provide instructions for handling retrieved information. For a domain-specific assistant, system instructions might tell the model to answer using supplied organizational content, avoid unsupported claims, or follow a particular output format. These instructions do not store embeddings or provision physical infrastructure. Their effectiveness should be validated through evaluation because language models can still produce unexpected results. System instructions are one component of application behavior and should be combined with retrieval, security, validation, monitoring, and other engineering controls.<\/span><\/p>\n

                                    Question 78. What is one potential risk of including too much retrieved context in a prompt?<\/b><\/p>\n

                                      \n
                                    1. <\/b> The model may receive irrelevant information and consume excessive context capacity<\/span><\/li>\n
                                    2. <\/b> The model automatically becomes more accurate<\/span><\/li>\n
                                    3. <\/b> The embedding model stops functioning permanently<\/span><\/li>\n
                                    4. <\/b> The application no longer needs retrieval<\/span><\/li>\n<\/ol>\n

                                      Correct Answer: 1. The model may receive irrelevant information and consume excessive context capacity<\/b><\/p>\n

                                      Explanation:<\/b><\/p>\n

                                      Providing excessive retrieved context can reduce the effectiveness of a RAG application. Large amounts of information may contain irrelevant or conflicting content, making it harder for the model to focus on the evidence most useful for answering the user’s question. Excessive context also consumes the model’s available token budget and can increase inference cost or latency. This is why chunking, ranking, filtering, and appropriate top-k selection are important parts of RAG design. More context is not automatically better. The goal is to provide sufficient, high-quality information that directly supports the task while avoiding unnecessary content. Retrieval strategies should therefore be evaluated for both relevance and completeness.<\/span><\/p>\n

                                      Question 79. What is the purpose of conversation history in a multi-turn RAG assistant?<\/b><\/p>\n

                                        \n
                                      1. <\/b> To provide relevant prior context for understanding follow-up requests<\/span><\/li>\n
                                      2. <\/b> To replace all external knowledge sources<\/span><\/li>\n
                                      3. <\/b> To permanently retrain the model after every message<\/span><\/li>\n
                                      4. <\/b> To eliminate authentication<\/span><\/li>\n<\/ol>\n

                                        Correct Answer: 1. To provide relevant prior context for understanding follow-up requests<\/b><\/p>\n

                                        Explanation:<\/b><\/p>\n

                                        Conversation history can help a multi-turn assistant understand follow-up questions that depend on previous exchanges. For example, a user may ask about a document and then refer to it as “that report” in a later message. Maintaining relevant history allows the application to interpret such references correctly. However, sending the entire conversation indefinitely may consume context-window capacity, so applications may summarize or selectively retain important information. Conversation history does not replace external knowledge sources when the assistant needs current or enterprise-specific information. It also does not retrain the model after every message. Effective conversation management balances continuity with context size, relevance, privacy, and performance.<\/span><\/p>\n

                                        Question 80. Which combination best supports a high-quality RAG application?<\/b><\/p>\n

                                          \n
                                        1. <\/b> Poor source data, unrestricted context, and no evaluation<\/span><\/li>\n
                                        2. <\/b> Accurate knowledge sources, effective retrieval, clear prompting, evaluation, and monitoring<\/span><\/li>\n
                                        3. <\/b> Random prompts and no access controls<\/span><\/li>\n
                                        4. <\/b> Maximum model temperature and unlimited document retrieval<\/span><\/li>\n<\/ol>\n

                                          Correct Answer: 2. Accurate knowledge sources, effective retrieval, clear prompting, evaluation, and monitoring<\/b><\/p>\n

                                          Explanation:<\/b><\/p>\n

                                          A high-quality RAG application depends on several components working together. Reliable source data provides a strong knowledge foundation, while effective chunking, embeddings, filtering, ranking, and retrieval help identify useful context. Clear prompts guide the generative model in using that context appropriately. Evaluation provides evidence about retrieval and generation quality, while monitoring helps identify problems after deployment. Security and governance should also be applied according to the application’s data and operational requirements. No single component guarantees high-quality results. In particular, increasing temperature or retrieving unlimited documents can introduce unwanted variability and irrelevant information. A systematic approach across data, retrieval, prompting, evaluation, and operations is therefore important.<\/span><\/p>\n

                                           <\/p>\n","protected":false},"excerpt":{"rendered":"

                                          View Full Databricks Certified Generative AI Engineer Associate Exam Dumps and Practice Test Dumps   Question 61. What is the primary purpose of prompt templates in a generative AI application? To provide a reusable structure for constructing model inputs To replace vector databases To increase physical storage capacity To manage network routing Correct Answer: 1. […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[1648,1647],"tags":[],"_links":{"self":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/18129"}],"collection":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/comments?post=18129"}],"version-history":[{"count":1,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/18129\/revisions"}],"predecessor-version":[{"id":18130,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/18129\/revisions\/18130"}],"wp:attachment":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/media?parent=18129"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/categories?post=18129"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/tags?post=18129"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}